Biometrisches akustisches schreibsystem und verfahren zur personenidentifikation und handschriftenerkennung mittels biometrischer daten
Abstract
This record has no abstract on file.
Term
Term ended
Projected expiry passed 4 September 2023, 3.1 years ago.
- Priority
- Filed
- Published
- Projected expiry
- Today
114 claims: 20 independent, 94 dependent
- 1Translation of claims of equivalent WO 2004023382 A1 1. Biometric acoustic writing system (1) with:(A) a pen housing (31) for performing hand-guided movements on a base (4);(B) at least one in the pen housing (3) integrated microphone (5) for the acoustic detection of sound signals caused by the hand-guided movements;(C) and a data processing unit (11) for calculating biometric data in dependence on the detected sound signals.
- 99th Biometric acoustic writing system according to one of the preceding claims, characterized in that the microphone (5) is arranged in an airborne sound chamber (7) provided in the pen housing (3).
- 2525th Biometric, acoustic writing system according to Anspru ch 27, characterized in that in addition a tilt sensor (40) is provided which detects the tendency of the writing pen (2) and thus the motor movement of the pen-guiding fingers when writing.
- 5151st Biometric writing system according to one of the preceding claims, characterized in that the detected sensor signal data via a data transmission path (18) from the pen housing (3) to a in a local computer (20) integrated data processing unit (19) are transmitted.
- 5252nd Biometric writing system according to one of the preceding claims, characterized in that in the pen housing (3) an encryption unit (17) for encrypting reference data of the sensor signal data is provided.
- 5454th Biometric writing system according to one of the preceding claims, characterized in that a data memory for storing biometric reference data, position data of the writing movement and of spoken information are provided.
- 5555th Biometric writing system according to one of the preceding claims, characterized in that the biometric reference data by the data processing unit (11) from the recorded during writing and speaking at least one word sound signal data from optical motion data, mechanical vibration and pressure data and tilt data are calculated and stored in a reference data memory ,
- 5757th Biometric writing system according to one of the preceding claims, characterized in that the reference data memory is a microchip of an identity card, a credit card, an authorization magnetic card or a memory unit of a computer or of the writing system (1).
- 6565th Biometric writing system according to one of the preceding claims, characterized in that the data processing unit (11, 19) compares the calculated current biometric data with the stored biometric reference data for their verification and identification.
- 7070th Biometric writing system according to one of the preceding claims, characterized in that with the stored biometric reference data of a person identified or verified via the handwritten input currently written, simple characters are reconstructed.
- 7171st A method of generating personal biometric reference data comprising the steps of:(a) acoustically detecting hand-held writing movements performed with a pen (2) on a pad (4) when a character writes a word, word or phrase, and generating corresponding sound signal data;(b) storing the generated sound signal data as a digital scarf1 time signal;(c) calculating frequency spectra as a spectrogram from the time-segmented sound signal data by means of fast Fourier transformation;(d) determining amplitude-time signals of selected frequencies for detecting the amplitude dynamics in the calculated spectrogram;(e) calculating an associated frequency spectrum from the selected amplitude time signals by means of fast Fourier transformation;(f) determining first biometric data from the sound and vibration intensity of the digital time signals by means of feature extraction;(g) determining second biometric data from the calculated spectrogram of the time-segmented sonic and vibrational time signals by means of feature extraction. (h) determining third current biometric data by means of feature extraction from the calculated frequency spectrum of the amplitude-time signals;(i) determining fourth current biometric data by means of feature extraction from acquired dynamic writing pressure, vibration and slope data;
- 8787th Method for verification and identification of a person with the following steps:(A) acoustically detecting hand-held writing movements performed with a pen (2) on a pad (4) in writing a character, word or phrase by the person and generating corresponding sound signal data;(b) generating writing pressure signal data for writing pressure and vibration signal data for vibrations transmitted from the writing pen to at least one pressure and vibration sensor;(c) storing the generated signal data as digital time signals;(d) calculating frequency spectra as a spectrogram from the stored time-segmented sound and vibration time signals by means of fast Fourier transformation;(e) determining amplitude-time signals of selected frequencies for detecting the amplitude dynamics in the spectrogram of the sound and vibration timing signals;(f) calculating an associated frequency spectrum from the selected amplitude-time signals by means of fast Fourier transformation;(g) determining first current biometric data from the sonic and vibrational intensity of the digital time signals using feature extraction;(h) determining second current biometric data from the and vibration time signals using feature extraction;(i) determining third current biometric data by feature extraction from the frequency spectra of the amplitude-time signals;(j) determining fourth current biometric data by means of feature extraction from the dynamic writing pressure data and;(k) comparing the current biometric data with stored biometric reference data of the person to verify that the current biometric data is substantially consistent with the stored biometric reference data of the individual;(1) comparing the current biometric data with stored biometric reference data of a plurality of individuals to identify whether the current biometric data is substantially consistent with the stored biometric reference data of one of the stored individuals.
- 9898th Method according to one of the preceding claims 87 to 96, characterized in that in a training phase reference feature vectors and in a working phase, the current feature vectors for corresponding characters, sketches or words from the acoustic writing signal data, the optical position data, the writing pressure data and voice data are determined and stored ;
- 9999th Method according to one of the preceding claims 87 to 96, characterized in that from feature vectors of the acquired signal data by means of statistical, connectionist and knowledge-based methods corresponding characters, picture elements, sketches, word segments, words or word sequences are determined.
- 100100th Method according to one of the preceding claims 87 to 96, characterized in that the acoustic handwriting recognition (ACR), the optical handwriting recognition (OCR) and the speech recognition integrated in one and the same writing system and the methods for image and text recognition are combined.
- 101One hundred and first Method according to one of the preceding claims 87 to 96, characterized in that the characteristics of the acoustic and optical data and printing data for the reconstruction of input characters, picture elements, words and texts are correlated and then stored as feature vectors.
- 102102nd Method according to one of the preceding claims 87 to 96, characterized in that from the correlated feature vectors of the acquired signal data by means of statistical, connectionist and knowledge-based methods according to characters, pixels, sketches, word segments, words or word sequences are determined;
- 114114th Use of the biometric writing system as a multifunctional computer input system of a virtual desk.
Independent claims20
181 paragraphs in 1 section, as filed
Translation of description of equivalent WO 2004023382 A1
Biometric acoustic writing system and method for personal identification and handwriting recognition using biometric data
The invention relates to a biometric system acoustic Cards are acoustically detected at the with an integrated in a pen device microphone projecting called when manually guided movement of the stylus sound signals and then processed to biometric data. The invention further relates to a method for personal identification and verification persons by means of a biometric writing system and a method for handwriting recognition and / or sketches recognition.
Currently, there are a wide range of competing biometric systems. The largest share of biometric systems currently account for systems that use finger lines as recognition, followed by systems for facial recognition and so-called iris scanner.
The increasing electronic networking in all areas of business life, the importance of electronic password and signature recognition is constantly increasing. Contracts are closed in a conventional manner by the two parties, by a signature is for certification. The signature serves as evidence and is specifically for those who does the signature. In contracts in the field of e-commerce is necessary that the contractor must provide an electronic signature allowing their identification, which must be verified for validity. Prerequisite for the widespread use of an electronic signature for verification of the identity of the parties is a high falsification security of performed electronic signature. Furthermore, the work signature must be detected with a high detection rate and transmitted to a remote computer Ü. Essential is the fact that the input of the electronic signature can be done for a large number of people without any special knowledge in a simple manner.
The previously known writing systems do not meet these requirements and exhibit some serious disadvantages.
There are pens known with accelerometers (for example,<sub>^</sub> Accelerometer-based pens of British Telecom munications) for receiving acceleration forces when writing with the pen. The disadvantage of this known stylus is in a comparatively low recognition rate in the signature verification because the only two-dimensional data space. In addition, the pen when writing must always be equally aligned, ie the angle between the pen and the base must be as constant as possible, so that a natural writing performance when bars of a signature is not guaranteed. In addition, an acceptable recognition rate is preceded by a minimum writing speed, otherwise the measured acceleration forces are too low.
A variety of writing systems recognize the handwritten text and characters by the writing pad. All such writing systems have the drawback that special writing pads with appropriate signal processing are required.
From US 4,513,437 a data input pin for signature verification is known. The pen contains a pen for visualization of the writing on a pad and piezoelectric pressure and acceleration sensors for detecting acceleration forces generated by transversal writing pen movement caused on the blotter. In addition, the pressure on the stylus in the axially parallel direction of the stylus is detected.
The disadvantage of the data input pen described in US 4,513,437 is that pronounced acceleration forces during writing are required. To capture sharp changes in direction and / or speed changes rectilinear writing movement for signal detection are necessary.
In the documents US 5 294 792A, US 5,333,209 A, US 6,208,755 Bl, DE 196 44 109 Al and DE 100 28 138 AI describes writing systems for detection of writing pressure dynamics by means of pressure sensors and write transaction data with a CCD sensor.
There are writing systems are known in which in a pen device for carrying out a pen hand-guided movements is integrated on a substrate. The pen device includes an optical sensor device for detecting the carried out with the stylus movements gungs- as movement or position data. In addition, in the writing pen device, a further pressure sensing device for detecting the forces occurring during the movements performed of the stylus is integrated. A data processing unit is used to calculate biometric data in response to the sensed motion data and the writing pressure data collected.
There are commercial writing instruments that are equipped with the sensor above techniques.
So describes the so-called. Anoto digital pen, a writing system that transmits the written text on a computer can. The stylus consists of a ballpoint pen with a built-in digital camera. A force sensor detected Animal T setting up the writing face on a special paper. The special paper is a paper with a printed dot pattern. The integrated digital camera of Anoto stylus performs the writing movements of the stylus over the dot pattern. The camera captures a plurality of images per second. Each image is compared to the next screen and an integrated processor calculates considering the pen posture change direction and the moving distance of the built-in pen writing shaft. The Anoto special paper are a number printed dots that are applied with ink based on carbon and have a distance of less than 1 mm each. Each Anoto- special paper has a different pattern of dots, so that the individual sites are distinguishable from each other. A light emitting diode (LED) illuminates the carbonaceous ink dots, which absorb long-wave infrared light and therefore can be recognized for the integrated digital camera within the Anoto- pin. By the Anoto special paper, the relative movement of the pin as well as the absolute position of the pen on the paper can be determined.
The disadvantage of the Anoto pen is that a special paper is needed and that the pin is predetermined attitude. In addition, the falsification of the signature verification is too low, there is no person-specific biometric data are determined. By Anoto- pen the individual writing pressure is not recorded on the write face.
In addition to the electronic signature verification is an increasing demand for electronic recording of handwritten texts and character. In conventional methods, handwriting recognition handheld writing movements are detected optically and the determined font data are then converted by the OCR process in ASCI-code data. Conventionally, in the recognition of a written word selects several possible word candidates. The number of possible word candidates is higher, the less clear is the handwriting and ever exist more similar words. From the various possible word candidates the most likely word candidate is then determined and output by a statistical analysis. Statistical analysis was carried out as in the automatic speech recognition means of so-called. Word sequence statistic programs that calculate the statistical probability of the occurrence of each word candidate from the foregoing recognized words. The most likely occurring word candidate is selected and then output. The disadvantage of such conventional methods in handwriting recognition is that the number of possible word candidates in many cases is very high, so that the detection rate is often very low.
Writing systems for signature verification, for personal identification or -Verifikation and for handwriting and speech recognition must have a very high falsification security and low error rates. Moreover, should the technical implementation not be too complicated to allow a wide range of applications.
It is therefore an object of the present invention to provide a multi-functional biometric acoustic writing system having low error rates and very high falsification security features, the widest possible acceptance possible and can be manufactured very inexpensively. This object is achieved by a biometric writing system having the features specified in claim. 1
The invention biometric writing system is preferably used as a multifunctional input system in biometrics for personal identification.
The writing system of the invention, the tablet a PC replaced as an input medium or supplement in a virtual desktop through normal writing paper. The writing system of the invention is also used in life science for diagnosis and therapy, in medicine and in psychology.
The invention relates to a biometric acoustic writing system with a Schreibstiftgehause to carry hand-held motion on a base, at least one integrated in the Schreibstiftgehause microphone for acoustic detection of acoustic signals, which are caused by the hand-operated movement, and with a data processing unit for calculating biometric data as a function of the detected sound signals.
The inventive biometric acoustic writing system allows the generation of person-specific biometric reference data, which are unique for the individual. This makes it possible to verify the identity of a person writing and identifying by comparison with stored reference data even.
The invention biometric writing system also enables handwriting recognition or sketches recognition for determining and issuing handwritten characters in digital form. The inventive biometric writing system preferably includes a data processing unit for reconstructing handwritten characters and texts from the detected sound signals.
In the Schreibstiftgehause a writing pin is preferably provided, which is placed on the base and guided on the base.
This offers the particular advantage that the person has a natural writing experience since the invention biometric writing system behaves like a normal pen device having an ink-leading pencil lead.
This is caused by the friction of the pen on the pad during the hand-held writing movement acoustic write noise is preferably transferred as structure-borne sound signal through the stylus and / or additionally as airborne sound signal via the ambient air through which is integrated in the Schreibstiftgehause microphone.
Since the write noise can not be transmitted solely through the ambient air, but also as a structure-borne sound signal is writing system the invention is also suitable for areas of application in acoustic noise environments.
The microphone is preferably mechanically coupled to the transmission of structure-borne sound signal with the stylus.
In this case, the microphone being mechanically coupled, in a particularly preferred embodiment for the transmission of structure-borne sound signal with a sound associated with the pen body. The sound body is preferably used as a resonator for be<sup>¬</sup> agreed eigenfrequencies designed.
In a preferred embodiment of the biometric writing system according to the invention the microphone is placed in a recess provided in the Schreibstiftgehause airborne sound chamber.
The air sound chamber is preferably designed as a resonator for certain natural frequencies.
The microphone is preferably surrounded with the resonators of a sound, which is provided for attenuation of ambient noise and only Ü transmits sound signals to the microphone via the stylus.
The air sound chamber can be coupled via a Schreibstiftgehause present in the housing opening with the ambient air.
The housing opening is preferably closed by means of a mechanical locking device for attenuation of ambient noise.
This offers the user the ability to suppress by closing the opening noise from the environment.
The microphone picks up with opened housing opening preferably caused by the hand-guided writing movement internal and external noise as write-borne and airborne sound signal and / or a product originating from a person speech signal.
The sound body for the external write noise is preferably formed by the blotter. The microphone with opened housing opening can optionally be replaced by an acoustically non-isolated second microphone in Schreibstiftgehause.
By the acoustic sensing of the speech signal in addition to the acoustic detection of the writing noise, the error rate can be significantly reduced in the detection of the person. By combining the analysis of the speech signal and the write noise a very high detection reliability of the biometric writing system is achieved.
The microphone converts the detected acoustic sound signals preferably into an electrical sound signal.
The electric sound signal is preferably nalog- by an A / D converter in the sound signal data for digital data processing converted by the data processing unit.
Here, the write noise and the speech signal are successively detected simultaneously or sequentially by the microphone.
The sound signal data generated are preferably be stored in a memory unit of the biometric writing system.
The blotter is preferably made of any paper.
The blotter is preferably a solid support having a specific surface roughness and hardness impressed. In a particularly preferred embodiment, the biometric acoustic writing system in the Schreibstiftgehause a speaker for playback of captured microphone signals for playback of stored biometric reference data and reproduction of spoken information is provided.
Preferably, in addition, a vessel of the person voice signal is acoustically detected and generates corresponding sound signal data for the speaker recognition or speech recognition.
The write noise itself can also be evaluated using the algorithms and methods of the speaker and speech recognition.
In a further preferred embodiment of the biometric, acoustic writing system according to the invention or acousto-mechanical write system, a second sensor means is further provided which Ü the writing dynamics over the static and dynamic pressure and the writing speed about the vibration of the stylus of the mounted on the substrate and hand-held writing instrument detected.
In a further preferred embodiment, the biometric acoustic writing system according to the invention or acousto-mechanical system of writing, at least three sensors for the simultaneous detection of occurring typists and vibrations of the writing shaft are provided in three directions.
In a further preferred embodiment, the biometric acoustic writing system according to the invention or acousto-mechanical writing system, an inclination sensor device is also provided, which determines the slope of the patch on the pad and hand-held writing instrument and thus detects mainly find their fingers when writing. In a preferred embodiment of the biometric writing system, an optical sensor device is additionally provided, detects the position data of the hand-held stylus motion across the image signals from the surface of the blotter.
In a further preferred embodiment of the biometric writing system further comprises a pressure sensor device is provided, which detects the static and dynamic pressure of the mounted on the support and hand-held stylus.
The microphone is preferably a Elektrett microphone, a piezoelectric microphone, a piezoresistive microphone or a capacitive microphone.
In the pen of the biometric writing system a replaceable pencil lead is preferably provided.
This pencil lead is preferably extensible by means of a switching device of the pen.
In a preferred embodiment of the biometric writing system according to the invention, the optical sensor device to an imaging optics for imaging the substrate surface and a converter means for converting the optical image signal into an electrical signal.
The imaging optics preferably consists of optical lenses and / or glass fibers.
In the transducer device is preferably a CCD camera or a photodiode array. Further, preferably a diode light source for illuminating the support surface is provided.
In a particularly preferred embodiment, a laser diode for illuminating the substrate upper additionally Smile Friend and integrated in the pin optical grating is provided.
The optical sensor device preferably detects the carried out with the pen movements by comparing the time-shifted recorded image sequences of the pad surface.
In an alternative embodiment, the optical sensor means detects the carried out by the stylus pen movements by comparing the offset in time, the recorded image sequences of the interference of the light reflected at the grating and the substrate surface the laser light.
The pencil lead is in a preferred embodiment in an extended state on the substrate by carrying out the hand-held movements a writing fluid from which increases the structuring of the substrate surface, so that the optical detection of the hand-guided movements facilitated by means of the optical sensor device and the ink image on the substrate a natural notation is caused.
The Dätenverarbeitungseinheit is preferably integrated in the Schreibstiftgehause or in an external receiving unit.
The external receiver unit is preferably a computer, a microcontroller, a cell phone, a credit card reader, a facsimile machine or a printer. The detected sound signals optical motion data and the detected multidimensional print data is preferably transmitted via a data transmission path of the Schreibstiftgehause to the integrated into external receiving unit data processing unit.
The inventive biometric writing system preferably comprises an encryption unit included in the Schreibstiftgehause for encryption of sound signal data, optical motion data and print data.
The data transmission path is preferably wire or wirelessly.
The inventive biometric writing system preferably comprises a data memory for storing biometric reference data of position data of the write and movement of spoken information.
The biometric reference data are preferably calculated by the data processing unit from the recorded in writing and speaking a word sound signal data from optical data movement and mechanical pressure data and stored in the reference data storage.
In the reference data storage is preferably a microchip of an identity card, a credit card to a Berechtigüngsmagnetkarte or to a storage unit of a computer or the biometric writing system.
In the written and spoken words and characters is preferably about pin codes, passwords, o- names to texts. The position data of the writing movement are preferably optical data movement and mechanical pressure data calculated from the sound signal data recorded when writing a word and stored in the data memory for handwriting recognition.
The reference data memory is preferably integrated into the Schreibstiftgehause.
The integrated in the external receiving unit data processing unit is preferably connected to a reference data storage.
The external receiver unit has a reading unit for reading a portable of biometric reference data in a preferred embodiment of the biometric writing system.
The external receiving unit is preferably connected via a data network to a database with a reference data memory.
When the data network is preferably the Internet.
In a preferred embodiment of the biometric writing system, the data processing unit compares the calculated dentifi current biometric data with the stored biometric reference data for their verification, and I-<sup>'</sup>cation.
The data processing unit generates an identification or verification indication signal in the substantial conformity of the current biometric data with the stored reference data. In a particularly preferred embodiment, the biometric writing system according to the invention the data processing unit detects at complete conformity of the current biometric data with the stored biometric reference data, the current biometric data as a pirated copy of the stored reference data, and generates a corresponding warning signal.
The data processing unit generates at deviation of the current biometric data with the stored biometric reference data a deviation indicative signal.
In a preferred embodiment of the biometric system of this writing has at least one actuator which is operated by generating the identification or verification indication signal.
The invention further provides a method for the generation of person-specific biometric reference data comprising the steps of:
Acoustic sensing handgefuhrten writing movements, which are carried out with a pen on a pad when writing a character, a word or phrase by a person, and generating corresponding sound signal data;
Save the sound signal data generated as a digital sound-time signal;
Calculating respective frequencies as a spectrogram of the sound signal zeitsegmentierten data using fast Fourier transform;
Determination of amplitude-time signals of selected frequencies to detect the amplitude dynamics of the calculated spectrogram; Calculating a corresponding frequency spectrum from the selected amplitude-time signals using fast Fourier riertransformation;
Determining first biometric data from the sound and vibration intensity of the digital time signals using feature extraction;
Determining second biometric data from the calculated spectrogram of zeitsegmentierten sound and vibration signals by time feature extraction.
Determining third current biometric data by means of feature extraction from the calculated frequency spectrum of the amplitude-time signals;
Determining fourth current biometric data by means of feature extraction of detected dynamic Schreibdruck-, vibration and inclination data.
In a preferred embodiment of the method for the generation of person-specific reference biometric data, the method steps are performed multiple times, and the respectively determined reference data statistically evaluated, wherein the evaluated data is encrypted as a person-specific reference data and subsequently stored.
In a preferred embodiment, the statistically evaluated, person-specific reference data in a microchip of an identity card, in a microchip of a credit card, in an authorization magnetic card or in the memory unit of a computer or a writing system are stored.
Preferably, in addition, a vessel of the person answering signal are acoustically detected and generates corresponding sound signal data. Furthermore, the writing forces occurring in the performed writing movement of the pen in at least one spatial direction are detected by pressure sensors, and generated according to at least one dimensional write print data in a preferred embodiment of the inventive method to generate person-specific biometric reference data.
Further, preferably the optically detected when writing a character, a word or character / word sequence with the pen on the pad handheld writing movements and generates corresponding write transaction data.
The biometric reference data are preferably determined from the sound intensity of the digital sound signal by means of feature extraction, from the calculated frequency spectrum of the sound time signal by means of feature extraction and from the produced in writing and speaking a word optical motion data and writing pressure data and in the references stored rerizdatenspeicher encrypted as person-specific biometric reference data.
It is preferably carried out a data reduction of the static and dynamic random motion data, the writing pressure data of the optical motion data and the sound signal data by means of Clusterbildüng.
From the data reduced cluster feature parameters are preferably extracted.
The generated biometric reference data are stored preferably inside or outside the pen.
The invention further provides a method of verification of a person comprising the steps of, acoustic sensing of hand-held writing movements, which are carried out with a pen on a pad when writing a character, a word or phrase by the person and generating corresponding sound signal data;
Generating writing pressure signal data for writing pressure and vibration signal data for vibrations, which are transmitted from the pen to at least one of pressure and vibration sensor;
Storing the signal data generated as a digital time signals; Calculating frequency spectra of a spectrogram of the stored zeitsegmentierten sound and vibration time signals using fast Fourier transform;
Determining amplitude-time signals of selected frequencies to detect the amplitude dynamics in the spectrogram of the sound and vibration time signals; Calculating a corresponding frequency spectrum from the selected amplitude time signals using fast Fourier transform.
Here, a vessel of the person answering signal are preferably also acoustically detected and generates corresponding sound signal data.
Here, a vessel of the person answering signal is preferably additionally acoustically detected and generates corresponding sound signal data.
Preferably, in addition, a vessel of the person answering signal is detected acoustically, and corresponding sound signal data are generated.
The hand-held while writing or sketching a character, a word or character / word sequence with the pen on the pad of writing is preferably additionally detected optically and generates corresponding position data.
Preferably, in addition, the forces occurring in the performed writing movement of the stylus in at least one spatial direction are detected by pressure sensors and generates appropriate, at least one-dimensional writing pressure data.
In a preferred embodiment of the method the mark, the word or phrase from the sequence of selected reference feature vectors from the position data and from the writing pressure data using statistical, connectionist and knowledge-based method is determined.
In a particularly preferred embodiment of the inventive method the determined character of the person carrying out for their control are optically and / or acoustically confirmed.
The generated person-specific reference data is preferably evaluated to determine psychological and / or physiological characteristics of the person.
The inventive biometric writing system is often used.
In a preferred embodiment, the biometric writing system is used as a computer input device.
In an alternative use of the system according to the invention writing is used as a voice input device, in particular as a voice recorder. The inventive biometric write system is preferably used to detect neuromotor disorders movement of a person.
The inventive biometric writing system is also preferably used as a therapy system to correct neuromotor movement disorders a person.
The inventive biometric writing system is also preferably used as graphological system for determining psychological / physiological characteristics of a person.
The inventive biometric writing system is preferably used as a training system for learning of writing.
The inventive biometric writing system is preferably used as a training system for learning a language.
Further are described preferred embodiments of the biometric writing system according to the invention and the method for personal identification, the persons verification, for the generation of biometric reference data and for handwriting and sketches detection with reference to the attached figures for explaining features essential to the invention.
Show it:
Figure 1 is a schematic illustration showing a preferred embodiment of the biometric writing system according to the invention. Figure 2 is a sectional view through a Schreibstiftgehause the biometric writing system according to the invention.
Fig. 3a-3e is a sectional view preferred by Schreibstiftgehause embodiments of the biometric writing system according to the invention;
Figure 4 is a Ablaufdiagra m explaining the handwriting recognition means of the biometric writing system according to the invention.
Figure 5 is a sketch to illustrate an optical or acoustic feedback for the writing person.
6 acoustically detected writing noises at different writing speed on cardboard and writing paper.
Figure 7 an acoustically detected sound time signal at the triple writing the letter "a".
Figure 8 is an acoustic sound locked-time signal at the triple letter of the letter "b".
Fig. 9 frequency signals while writing the character "a" or "b" in the course of time;
Figure 10 is an amplitude-frequency signal when writing an associated letter "a.";
Fig. 11 dynamic behavior of some frequency lines while writing the letter "a";
Fig. 12 dynamic behavior of some frequency lines while writing the letter "b"; FIG. 1 shows a schematic block diagram of the biometric writing system according to the invention 1. The invention biometric writing system 1 includes a stylus 2, which is integrated into a Schreibstiftgehause. 3 The stylus 2 is used to carry hand-held motion on a base 4. The base 4 may be any document with any surface. Preferably, in the pad 4 to paper. The writing pad 4 is preferably a solid support having a specific surface roughness and hardness impressed.
The stylus 2 preferably contains a replaceable pencil lead, which can be extended by means of a not shown mechanical switching device from the pen. 2 The Schreibstiftgehause 3 is used to perform hand-guided movements on the pad 4. In this case, a writing fluid is the presentation of a normal writing feeling by the stylus 2 and the pencil lead delivered to the pad 4 to the optical feedback for the writing person. When placing the pen 2 on the pad and its movement on the surface arise writing noises. In the Schreibstiftgehause 3 at least one microphone 5 is provided for acoustic detection of acoustic signals in the writing system 1 according to the invention. The microphone 5 detects the acoustic signals that are caused by the hand-held movements, that is caused by the writing movement write noise. Provoked by the friction of the pen 2 on the pad 4 during the hand-held writing movement acoustic writing noises are transmitted on the one hand as a structure-borne sound signal via the pen 2 and on the other hand as airborne noise signal via the ambient air to the microphone. 5 Here the microphone 5 for transmission of structure-borne sound signal is mechanically the pen 2 coupled. When in Fig. 1 illustrated embodiment, the microphone is 5 mechanically coupled to the transmission of structure-borne sound signal to a connected to the pen 2 sound body 6. The sound body 6 is preferably designed as a resonator for certain natural frequencies.
As shown in FIG. 1, there is the microphone 5 in an opening provided in the magazine pin housing 3 airborne sound chamber 7. The airborne sound chamber 7 is preferably designed as a resonator for specific natural frequencies. The microphone 5 is surrounded by a sound insulation 8a, 8b, which is provided for attenuation of ambient noise and lets through only sound signals from the sound body 6 and the air switching chamber. 7 The airborne sound chamber 7 is coupled via an existing in the Schreibstiftgehause 3 housing opening with the ambient air. The housing opening is preferably closeable for attenuation of ambient noise by means of a mechanical locking device. 9 The microphone 5 detects with opened housing opening, the caused by the hand-guided writing movement internal and external write noise as body- and airborne sound signal and an additional derived by a person speech signal. The microphone 5 converts the detected acoustic sound signals into an electrical sound signal. The electric sound signal is converted by an analog / digital converter in the sound signal data for digital data processing. The digital signal data is supplied via a signal line 10 of a data processing unit 11 for further data processing. The write noise generated when writing and optionally additionally received by the person answering signal either simultaneously or consecutively detected by the microphone 5 and converted into corresponding sound signal data. The data processing unit 11 supplied sound signal data is preferably in a designated Speicherein- unit 12, which is connected to the data processing unit 11 via lines 13, are stored.
In addition to the microphone 5 as an acoustic sensor, the illustrated in Fig. 1, the invention biometric writing system 1 preferably additionally an optical sensor device 14 that detects the position data of the hand-held stylus motion across the image signals from the surface of the blotter. 4 The optical sensor device is also connected via a signal line 15 to the data processing unit eleventh
The microphone 5 can be, for example, an electron trett microphone, a piezoelectric microphone, a piezoresistive microphone or a capacitive microphone. The optical sensor device 14 preferably includes an imaging optics for imaging the ünterlagenoberflache and a converter means for converting the optical image signal into an electrical signal. The electrical signal is converted by an analog / digital converter into position data and motion data and output via the signal line 15 to the data processing unit eleventh The imaging optical system of the optical sensor device 14 is preferably made of optical lenses and / or glass fibers. In the transducer device may be a CCD camera or a photodiode array. Preferably, in addition, a diode light source for illuminating the support surface is provided. Further, a laser diode to illuminate an integrated in the pen optical grating is provided. The optical sensor device 14 detects the conducted by the stylus pen 2 movements by comparing the time-shifted recorded image sequences showing the surface of the pad. Alternatively, 14 can be also adapted the optical sensor means that it the offset carried out with the pen 2 movements by comparing time captured image sequences of interference of at. the grid and the back-up surface reflected laser light detected.
The pencil lead is in the extended condition to the backing 4 in executing the hand-held movements a writing fluid from which increases the structuring of the surface, so that the optical detection of the hand-guided movement is facilitated by means of the optical sensor device fourteenth Further effects on the ink image on the document with the person writing a natural notation.
The data processing unit 11, as shown in Fig. 1, integrated in the Schreibstiftgehause third In an alternative embodiment, the data processing unit may also be integrated in an external receiving unit. Such receiving unit is for example a computer, a microcontroller, a cell phone, a credit card reader, a facsimile machine or a printer.
The detected by the microphone 5 sound signal data and detected by the optical sensor device 14 movement data, provided that data processing unit externally located in a local computer, transmitted via a data transmission path 18 of the Schreibstiftgehause 3 to this local computer 20th The data processing unit 11 is, as can be seen in Fig. 1, connected via a signal line 16 to a downstream encryption unit 17. The encryption unit 17 is used to encrypt the output from the data processing unit 11 within the pen housing 1 sound data and the detected optical motion data. The encrypted data is on the data transmission path 18 which is wired or wireless, to a data processing unit 19 within the local computer 20 transmitted. The local computer 20 has, besides the data processing unit 19 via a memory 21 and a display 22 respectively through signal lines 23, 24 are connected to the data processing unit 19th The data processing unit 19 of the local computer 20 is further connected via lines 25 to a reading unit 26th Moreover, the data processing unit 19 of the local computer 20 via control lines 27 controls an actuator 28, for example, a security door. The data processing unit 19 is connected via a line 29 to a data network 30 to which an external database 31 is connected. In the data network 30 is preferably the Internet.
The illustrated in Fig. 1 biometric writing system 1 includes an acoustic sensor device 5 and an optical sensor means 14. In a preferred embodiment, the invention biometric writing system 1, in addition to a pressure sensor device which detects the static and dynamic pressure of the mounted on the substrate 4 hand-held stylus and corresponding write print data to the data processing unit 11 outputs. The detected sound signal data optical motion data and palpate multidimensional print data are transmitted from the data processing unit 11 after appropriate encryption by the encryption unit 17 via the data transmission path 18 of the Schreibstiftgehause 3 to the integrated into the local computer 20 data processing unit 19 Ü. From the detected sound signal data, the optical motion data, and from the mechanical write print data biometric reference data by the data processing unit, which is located within o- outside of the stylus housing 3, calculated in a reference data memory, for example in the memory 12 or the memory 21, is stored. In the reference data storage may be in alternative embodiments to a memory within a Identity card, a credit card would constitute an authorization magnetic card or a memory unit of any computer. The written words and characters whose Schreibgerausche be acoustically detected by the microphone 5 and output as sound signal data are particularly spelled passcodes, passwords, names or handwritten text. The written by the person character can be spoken at the same time by the person and are recognized as a speech signal also by the microphone. 5
As provided for in the local computer 20 reading unit 26 is used for example for reading a portable storage medium for biometric reference data. In the data processing unit 31 calculated current biometric data with stored reference biometric data for verification and identification can be compared. In the substantial conformity of the current biometric data with the stored reference data, the data processing unit 11 generates and 19, an identification or verification indication signal. Here, the data processing unit 19 to generate the identification or verification indication signal, an actuator, such as a security Stur, operated.
Detects the data processing unit 11 or 19 is a fully stan-ended match the current biometric data with stored biometric reference data, the current biometric reference data by the data processing unit 11, 19 identified as pirated copy of the stored reference data and a corresponding warning signal is generated.
The saved person-specific biometric reference data is first generated by the inventive biometric writing system 1 for later comparisons. The generation of person-specific biometric reference data is as follows. First, the hand-held be writable are movements, which are carried out with the pen 2 on the pad when writing a character, a word or phrase by a person acoustically detected, that the write noise is detected acoustically by the microphone. 5 The microphone 5 is generated by analog / digital conversion of sound signal data, which are transferred to the data processing unit eleventh The sonic signal data is stored as digital sound time signal and from the stored time sound signal an associated frequency spectrum is calculated by means of Fourier transformation. Using feature extraction can be determined from the sound intensity of the stored digital sound signal time first reference biometric data. second reference biometric data are determined by means of feature extraction from the calculated frequency spectrum of the sound time signal. These process steps are preferably carried out several times, and the relevant benchmark data are statistically evaluated, the evaluated data are stored encrypted as person-specific reference data. The statistically evaluated, person-specific reference data are preferably stored in a microchip of an identity card, a microchip of a credit card, in an authorization magnetic card or in the memory unit of a computer or writing system. In addition to the write noise the microphone 5 can capture a product originating from the person answering signal acoustically and make appropriate sound signal data to the data processing unit 11th Preferably, the hand-guided writing movements are detected optically and generates corresponding write transaction data, which are sent to the data processing unit 11 in addition by the optical sensor device 14th Furthermore, preferably the typists occurring in the performed writing movement of the stylus 2 are recorded via pressure sensors, not shown, and submitted in accordance with at least one-dimensional writing pressure data to the data processing unit 11th The Datenverar- processing unit 11 determines the reference biometric data from the sound intensity of the digital sound signal by means of feature extraction, from the calculated frequency spectrum of the sound time signal by means of feature extraction and from the produced in writing and speaking a word optical motion data and writing pressure data generated. Is preferably followed by a data reduction of the static and dynamic random motion data, the writing pressure data of the optical motion data and the sound signal data by means of clustering. From the data reduced cluster feature parameters are then extracted.
Fig. 2 shows a preferred embodiment of a stylus pen 2 of the biometric writing system according to the invention 1. The illustrated in Fig. 2 Schreibstiftgehause 3 is for example an ordinary ballpoint pen casing. In the Schreibstiftgehause 3 is a pen 2 with integrated pencil lead. The stylus 2 is mechanically connected to a sound body 6 for transmission of structure-borne sound signals. The recording of the write noise by the microphone 5 via the pen 2 and the sound body 6. In an alternative embodiment of the stylus 2 itself can serve as a sounding body. In addition to the write noise the microphone 5 can acoustically detect a product originating from a person speech signal when the air switch chamber 7 or the acoustic window is connected by opening a closure device with the environment. The detection of the writing of the speech signal and noise can be effected by the microphone 5 at the same time or in sequence. When in Fig. 2 illustrated preferred embodiment, the microphone 5 is a Elektrett microphone. In the embodiment shown in Fig. 2, a loudspeaker 32 is additionally provided in the Schreibstiftgehause third The speaker 32 is used to playback captured microphone signals, and moreover, the speaker can 32 for reproduction stored biometric reference data and are used for reproduction of spoken information. At the sound<sup>¬</sup> speaker 32 it is preferably a miniaturized conventional speakers.
When writing in the movement of the pen 2 on the support 4, is formed an acoustic signal or a write Schreibgerausch. The pen 2 may consist of a ballpoint pen, an ink pen or a rod with a defined peak. The backing 4 consists for example of paper or a pad with natural and additionally imprinted surface roughness and hardness. The surface roughness of the support 4 leads when writing of the stylus 2 to statistically excited, forced vibrations of the pad and the pen, and therefore leads to statistical Schreibkratz- and Schabgerausche. This airborne sound waves and / or Korperschallwellen caused and transmitted to the microphone 5 within the Schreibschriftgehau- ses 3 depending on the design of the writing EXPOSED and documents. The pin 2 to the housing 3 as a sounding body and the airborne sound chamber 7 in the Schreibstiftgehause 3 are optimal according to the laws of acoustics are interpretations. The sound body 6 and the air sound chamber 7 are preferably set as resonators having certain frequencies. Here preferably frequencies are used that have write access to the most dynamic region in the Amplitudenan- ation.
When writing of the pen 2 is the microphone 5, a sound time signal which is well above the noise signal and the sound intensity is sensitively dependent on the writing speed and a very small part of the different roughness of the writing pad and the different average contact pressure of the writing lead. 2 The sound intensity of the recorded sound time signal increases with the writing speed. In the same writing speed the differences in intensity at different writing paper surfaces are negligible. The frequency spectrum of the captured sound time signal consists of a continuous noise spectrum which is significantly overshadowed by the characteristic frequency lines. The frequency lines are mainly associated with frequencies below 2 kHz. The amplitudes of these frequency lines take at the same frequency position significantly with the writing speed. The amplitudes of the frequency lines increase with increasing surface roughness of the writing pad 4 to a small extent. The frequency positions remain unchanged. In the same writing speed, the amplitudes are almost independent of the used blotter 4. The air silenced microphone 5 is insensitive to sound events in the area. If the Schreibstiftgehause 3 is completely closed, the air acoustic chamber has 7 no connection with the ambient air. With a closed air chamber 7 have ambient sounds, especially spoken words or engine noise, while writing no effect on the recorded sound time signal. In this case, the structure-borne sound generated on the writing lead is only picked up by the microphone. 5
The biometric features of the person writing dynamics are created from the temporal change of the acoustic microphone signals and / or the amplitudes of selected spectral lines' averages. Since for the generation of biometric reference data, only the dynamics of the received sound signals is evaluated, the influence of different roughnesses of the base 4 and the influence of manufacturing variations in the production of the writing instrument is negligible. Variations in the writing dynamics are therefore personenbedingt or person-specific and can therefore be evaluated for identification and verification of the person. The a- kustisch detected by the microphone 5 sound time signal must be opposed Writing pressure signals representing the writing shaft pressure in different directions, regardless of how the pen 2 is held on the base. 4 The sound signal produced by the microphone 5 does not change when the Schreibstiftgehause 3 is held rotated differently about the longitudinal axis. The writing implement shown in FIG. 2 can be dimensioned with respect to the geometry, the material, the layer structure, the attenuation and the acoustic coupling between the sound source and the microphone 5 that can be optimal, acoustically less susceptible to failure biometric features for authentication methods generate. In the Schreibstiftgehause 3 an acoustic window or an opening is provided, which can be opened by pressing a button by the person for voice recording. To this end a relatively small opening in the Schreibstiftgehause 3 sufficient to accommodate the air-borne sound during speech through the microphone 5 with. To capture the pure acoustic signal or write Schreibgerau- cal the Gehausefenster is closed by the person.
FIG. 3a shows a further preferred embodiment of the achieved findungsgemaßen biometric acoustic writing system 1. In the case shown in Fig. 3a illustrated, particularly preferred embodiment, additionally another sensor devices are in the Schreibstiftgehause 3a next to the microphone 5 is provided for the acoustic detection of Schreibgehausen. The m Fig. 3 represented biometric pen contains additionally suddenly an optical sensor device 33. The optical sensor device 33 detects the position data of the handgefuhrten pen movement on image signals from the surface of the desk pad 4. The optical sensor device 33 contains an imaging optical system for imaging the pad surface and a converter device for converting the optical image signal into an electric signal. The imaging optics includes an optical lens 33a. In the lens system, it may be glass fiber bundle or GRIN lenses. The conversion means 33c of the optical sensor device 33 is for example a CCD camera with an integrated microprocessor. Alternatively, a photodiode array can be used. Preferably, in addition, a diode light source for illuminating the support surface is provided. In an alternative embodiment the laser diode used in addition to the illumination of an integrated optical grating in the pin.
The optical sensor device 33 detects the conducted by the stylus pen 2 movements by comparing the time-shifted recorded image sequences of the pad surface. In an alternative embodiment, the optical sensor means detects the carried out by the stylus pen 2 movements by comparing the time-shifted recorded image sequences of the interference of the light reflected at the grating and the substrate surface the laser light. The pencil lead 2 are in the extended condition to the backing 4 in executing the hand-held movements from a writing fluid, which increases the structuring of the substrate surface. Thereby, the optical detection of ha'ndge- is led movements by means of the optical sensor device 33 easier, and the ink typeface a person receives feedback, whereby the natural spelling is facilitated.
The illustrated in Fig. 3a first preferred embodiment of the biometric writing system 1 according to the invention contains in addition to the microphone 5 for detecting an acoustic write noise and optical sensor means 33 for the generation of position or movement data in addition, a pressure sensor 34 for detecting the writing pressure. The pressure sensors 34 consists for example of strain gauges DMS or force-sensitive resistors FSR and piezo sensors 34a, 34b, 34c, 34d, which detect the static and dynamic pressure of the mounted on the substrate 4 and hand-held stylus 2 sensors. The emitted by the microphone 5 Sound signal data outputted from the optical sensor means 33 transaction data and the detected multidimensional writing pressure data of an integrated data processing unit 11a, 11b is supplied inside the stylus housing 3 and processed to biometric data. The detected biometric data is preferably stored in a data memory 12 within the pen housing. 3 The current detected biometric data is then preferably encrypted by an encryption unit 17 and transmitted to a remote computer 20 via a data transmission path 18th
In the embodiment shown in FIG. 3a can be 7 to switch between voice and writing sound recording by moving a sound 8b within the pen housing 3 the acoustic window. In a noisy environment can be suppressed, for example by closing the acoustic window 7 the disturbing exterior noise. To increase the reliability of detection, the person can open the acoustic window and for example, a written password simultaneously or sequentially speak further.
Figure 3b shows a second preferred embodiment of the biometric writing system according to the invention 1. The data processing unit 11 is powered by a rechargeable battery 37 with electricity. At the top end of the pencil lead 2 a pressure sensor device 34b is provided. Further, a spring 38 is provided. In the pressure sensor 34b is preferably a piezoelectric pressure sensor. Figure 3c shows a third preferred embodiment of the writing system of the invention 1. The writing system 1 is rotationally symmetrical and consists of a pencil lead 2, a microphone 5 with bracket, a sound insulation 8 and a piezoelectric pressure sensor. 3
3d shows a fourth preferred embodiment of the writing system 1 according to the Schreibstiftgehause 3 is a fingerprint sensor 39 is provided. The fingerprint sensor 39 detects biometric data of the person writing about physical characteristics. The fingerprint sensor is preferably a capacitive fingerprint sensor. In the Schreibstiftgehause 3 the pen 2 is supported by a spring steel bracket 42nd The stylus 2 has two microphones 5a, 5b. Within one provided in the Schreibstiftgehause 3 body 49 a thinner microphone 5a is provided, which is located near one side of 41st The side 41 serves as an acoustic oscillator and is connected to the pencil lead. 2 An outer microphone 5b serves for receiving the ambient sound. The writing system 1 in the embodiment shown in Figure 3d further includes a tilt sensor 40th The tilt sensor 40 detects when writing the inclination of the stylus 2. This is the motorized movement of the pin leading finger detected. The data processing unit 11 is powered by a rechargeable battery 37 with electricity. At the upper end of the stylus 2, is also located a pressure sensor 34d. Other pressure sensors 34a, 34b, 34c are provided on the spring steel bracket 42nd The spring steel bracket 42 serves as a pencil lead holder and as deformers the strain gages 34a, 34b, 34c, which are preferably mounted on the spring steel bracket 42nd
Figure 3e shows a fifth preferred embodiment of the biometric writing system according to the invention 1. The writing system 1 includes an optical sensor means 48 which simultaneously the static and dynamic pressure and the oscillation of the mounted on the substrate 4 handheld stylus pen 2 detects in three spatial directions. This optical sensor device 48 includes photodetectors for detecting the movement of a diode light source 44. The diode light source 44 is fixed to the pen 2. The diode light source 44 passes through the deflection and vibration of the stylus 2, wherein the output from the diode light source 44, light rays are emitted as the light source signal to the photo detectors of the optical sensor device 40th The signals emitted by the diode light source 44, light beams reach the one he Strahlentei 47 and from there to a lens 46. The lens 46 preferably has a partially mirrored underside. The optical sensor device 40 further includes a four-quadrant photodetector 50, and a simple photodetector 45. The four-quadrant photodetector 50 is mounted behind the lens 46th The four-quadrant photodetector 50 detects the displacement and the vibration of the light source 44 in mutually orthogonal X-, Y-directions. The simple photodetector 45 detects the displacement and vibration of the light source 44 in the Z to the X, Y-direction perpendicular thereto. From the signals of the four quadrant photo-detector 50 and the simple photodetector 45 are in three dimensions about the deflection, the dynamics of the writing pressure and the vibration of the dynamics of the writing speed can be determined. The signals of the photodetectors 45, 50 are the data processing unit supplied 11 and evaluated there. In the illustrated embodiment in Figure 3e the pen 2 is supported by a writing cartridge holder 43rd Further, a spring 38 is provided. The figure of the optical sensor device 40 consists of the beam splitter 47, the optical lens 46 and a partially reflecting aperture.
The illustrated in FIGS. 1 to 3 biometric acoustic writing systems 1 are versatile. The Biomet innovative writing system 1 can be used as a computer input device. In addition, the biometric writing system 1 is suitable as a voice input device, especially as a voice recorder.
Another possible application is the identification and verification of a person writing by comparing currently obtained biometric data with abgespei<sup>¬</sup> sured reference biometric data of individuals. During verification of a person, have to check whether, for example, the given signature actually originates from the right person. For this purpose, the currently generated biometric data with stored reference data of this person are compared. A prerequisite is that the first person-specific biometric reference data of this person will be generated. For this, first be hand-held Schreibbe<sup>¬</sup> movements that are performed by that person under supervision, acoustically recorded. The person carrying signs, for example, under supervision with their signatures, and the microphone 5 in the writing instrument detects the sound signals occurring the write noise. The write noise recorded are output as acoustic signal data to the data processing unit eleventh The sonic signal data is stored temporarily in a memory 12 as digital sound time signals. Subsequently, by the data processing unit 11, a calculation of the associated frequency spectrum, for example by means of Fourier transformation. first biometric reference data are determined by means of feature extraction from the stored digital signal switching time. second reference biometric data from the calculated and stored frequency spectrum of the signal switching time also means feature extraction determined. These method steps may be performed several times, preferably, and each specific biometric reference data are statistically analyzed. The evaluated data then encrypted as a person-specific reference data and stored. The generated reference biometric data can then be used for verification, and personal identification persons in a variety of applications. The generated specific personal reference data are stored for example in a microchip of an identity card, in a microchip of a credit card, a magnetic card authorization or in any storage unit of a computer. When generating the biometric reference data derived from the person answering signal is preferably additionally acoustically detected and generates corresponding sound signal data. For example, the person concerned can approach a bank or government agency provide a signature under the supervision and simultaneously or subsequently speak their own names. To further increase the safety and the write forces occurring during the signature performance also can be detected and evaluated by means of sensors of pressure. A further increase in security is obtained by additional optical detection of the hand-held writing movement, that is, an optical detection of position data of the rendered signature. The pen illustrated in FIGS. 3a-3e detected by sensors, a variety of personal-specific features, especially the typeface this signature, occurring during the writing of the signature static and dynamic writing pressure, the write noise generated when writing the signature and also the writing of the person spoken word, for example, their name or a password. All this data is processed and stored .. The detection reliability and the falsification security is very high, since different person-specific features are detected by sensors and evaluated for biometric reference data. For verification of a person generated by the writing system current biometric data with the previously generated and stored biometric be see references of the person compared to determine whether the current biometric data with stored biometric reference data of the person broadly consistent. In a wholly owned agreement is, however, believed to be a pirated copy of the stored reference data, and the biometric writing system 1 according to the invention emits a corresponding warning signal.
In the case of personal identification, the current biometric data with stored biometric reference data of a plurality of persons are compared, and the person selects the reference biometric data with the current biometric data are best matched.
In addition to the persons verification and personal identification, the inventive biometric writing system can also be used for handwriting recognition and / or sketches recognition additionally. Fig. 4 schematically shows the steps for handwriting recognition. The Schreibgerausche occurring during writing of the pen 2 on the base 3 when writing or sketching a character, a word or phrase to be acoustically detected by the microphone 5 and converted into corresponding sound signal data. After an acoustic signal processing is an encoding. The sound signal data generated are temporarily stored as a digital sound time signal, and then the associated frequency spectrum of the sound signal data is calculated by means of Fourier transformation.
In a further step first current acoustic data from the acoustic intensity of the digital sound time signal are determined by feature extraction for data reduction. current second acoustic data from the calculated frequency spectrum of the sound signal time also means feature extraction determined. This step also serves mainly to Data reduction. From the acoustic data finally a sequence of feature vectors are calculated, and this calculated feature vectors are compared with stored person-specific reference feature vectors in a pattern matching. There is then a sequence of reference feature vectors which have from respective feature vectors of the smallest deviation is selected. From the sequence of selected reference feature vectors, the written characters, the written word or the written word sequence are determined by statistical, konnek- tionistischer and knowledge-based methods.
The statistical method is, for example, hidden Markov models for the detection of a larger vocabulary. The connectionist method primarily neural network models for the recognition of a limited vocabulary are used. In the knowledge-based method is, for example, fuzzy logic models. The reference feature vectors are person-specific generated and stored in the position of the writing system specific to the system and prior to use of the system by repeatedly handwritten letter of a training text. Over extended periods of use the reference feature vectors of the person to be updated adaptively and automatically re-written characters and common. All methods are based on acoustic models of characters, letters, syllables, and these words of the vocabulary of a language in phonetics. The phonetics of the vocabulary of a language is by phonetic elements, ie short, phonetic units (allophones) formed the write Gerau cal. Each vocabulary word is stored as a sequence of allophones. The acoustic models determined from the vocabulary phonetically more word candidates that belong with the highest probability to the acoustic symbol sequence. The probability of each word is determined on the HMM (hidden Markov models) or neuro-fuzzy systems. The most probable word candidate is determined by a threshold decision or a word sequence statistics model or language model. The determined character is then output, for example as formatted text or handwritten text. To increase the detection reliability of the written character or word is preferably additionally pronounced by the person recorded acoustically by the microphone. In addition, the position data of the hand-guided writing movement are detected optically. A further increase in security is achieved by detecting the writing pressure data occurring during the writing of the word.
The data generated by the inventive writing system 1 person-specific reference data can be analyzed to determine psychological and or physiological characteristics of the person. In particular, the occurrence neuromotor disorders movement of a person can be detected by means of the invention biometric writing system first The inventive biometric writing system is therefore suitable for the diagnosis of psychological and / or physiological characteristics, diagnostics, neuromotor movement disorders of the person.
Fig. 5 shows a particularly preferred embodiment of the biometric writing system 1 according to the invention with a optoa- kustischen feedback for the writing person. From the pen 3, the obtained biometric data is transmitted to a local computer 20, the 35b of the person writing on a screen 36a and a speaker are connected via lines 35a, 36b, an opto acoustic feedback. In this configuration, according to the invention biometric writing system 1 is suitable as a system for the treatment of movement disorders the person. Due to the opto-acoustic feedback the person may have a neuromotor movement disorder can learn to write again. Another interesting field of application is the use of the biometric writing system 1, as shown in Fig. 5, as a training system for learning writing. In particular, the invention biometric writing system 1 is suitable as a training system for learning a language, especially a foreign language. The learning could be for example write a word on the pad 4, and the built-in speakers Schreibstiftgehause 3 gives the word in the correct pronunciation of a foreign language. Through a learning program for students or people with coordination problems, for example, stroke patients can preset numbers or characters that are mounted on the substrate 4, are traced by the person with the pen. For example, a student a predetermined number or a mark on the pad 4 and trace receives an opto-acoustic feedback. For people with coordination problems as a documentation of their learning progress is possible by the writing print version. For children it is by means of the inventive biometric writing system 1, as shown in Fig. 5, it is possible to learn to write the alphabet in a simple manner. In addition, production exercises Bewusstma- monitoring exercises and imitation exercises possible.
Fig. Figure 6 shows the behavior of the writing system 1 of the invention when writing to different documents 4. FIGS. 6a, 6b show the time recorded sound signal at different writing speeds on a board surface and on a writing paper block. Here in the example shown, a line with increasing speed on the carton and the writing paper block is drawn. The fi g. 6c, 6d show the associated frequency spectra. The FIG. 7 shows a time sound signal that is generated at the triple writing the letter "a" with the letter of the pen on a pad by the inventive biometric writing system 1. As can be seen FIG. 7, the sound time signal is very good repeatable and reproducible, so that the letter "a" can be specifically recognized.
Fig. 8 also shows a time sound signal, which is generated at the triple writing of a letter "b". Also, the letter "b" has a characteristic sound time waveform.
Fig. 9a shows amplitude frequency signals over time for the letter "a", and Fig. 9B shows amplitude frequency signals for writing of the letter "b" in the course of time.
Figs. 10a, 10b show in principle how the Amplitudenfre- quenzlinien itself over time dynamically when you write a letter, for example, a letter "a" change. At the time tO the writing person sets the pin on (starting point AP) and performs the writing movement. The existing at the time tO spectrum changes dynamically over time. Fig. 10b shows the change of the spectrum at the time tO to time tl.
The Fig. 11 show the dynamic behavior of some frequency lines when writing a letter "a". In the example shown, the best representation of the letter "a" results for the frequency line f = 86.133 Hertz.
The FIG. 12 show the dynamic behavior of some frequency lines when writing a letter "b". The best representation of the letter "b" provides the frequency line f = 86, 133 Hertz. The inventive acoustic writing system 1 can be configured such that it detects the frequency lines, the handwritten characters in the largest possible dynamic range with large amplitude values.
The present invention biometric acoustic writing system 1 has a plurality of applications. It can on the one hand be used as a biometric authentication system for personal identification and verification. The writing system 1 of the invention is suitable as a computer input device as a substitute for a computer mouse or keyboard shortcuts, especially for password entry. Furthermore, the inventive biometric writing system 1 can be used for electronic voice and handwriting recognition. Another major field of application is in the field of biometric diagnosis and therapy. Finally, the invention biometric writing system 1 can for training purposes, for example be used for learning writing, foreign languages or to troubleshoot incoordination. Important areas of application of the writing system of the invention will become apparent in the area of e-commerce, particularly on the Internet, such as in banking, retail and credit card being. The inventive biometric writing system 1 may also be used as a biometric system for Personenzugangs- and security control in all security areas of high corruption safety and low error rates. Other applications exist in forensic applications in forensics, border management or the police. In addition, the inventive biometric writing system 1 for communication between man and machine can be used, particularly as a voice recorder, or to verbal and handwritten machine control. The multi-functional biometric writing system 1 is also suitable as an input unit of a future virtual desktop, in which augmented or replaced, for example on the pen tablet a PC through the normal writing paper as an input medium with the input options invention.
The inventive biometric writing system 1, as shown in Fig. 1 to 3, can be manufactured at low cost with relatively little technical effort, so that open wide application area.
LIST OF REFERENCE NUMBERS
1 biometric acoustic writing system
2 pen
3 Schreibstiftgehause
4 pad 5a, 5b Microphones
6 soundBody
7 air acoustic chamber 8, 8a, 8b soundproofing
9 closure device
10 signal line
11 data processing unit
12 storage unit
13 lines
14 optical sensor device
15 signal lines
16 signal lines
17 encryptor
18 data link
19 data processing unit 0 local computer store 1 2 display 3 line 4 line 5 line 6 reading unit 7 control line 8 Actuator 9 line 0 data network 1 Database 2 Loudspeaker 3 second optical sensor device Pressure sensor device line acoustic feedback devices battery spring fingerprint sensor tilt sensor string spring steel bracket writing shaft mount LED photodetector lens beam Reiber first optical sensor device body four-quadrant photodetector
Every citation, both waysCites: the store holds 3 of 4
| Document | Relation | Office | Cited during |
|---|---|---|---|
| US4234868A | Cites | United States of America | Examiner |
| US5850058A | Cites | United States of America | Examiner |
| US6307956B1 | Cites | United States of America | Examiner |
| See also references of WO 2004023382A1 | Non-patent | – | Examiner |
7 members in 5 offices
Priority claims14
| Document | Office | Kind | Date |
|---|---|---|---|
| 10241328 | Germany | A | |
| 10241328 | Germany | A | |
| 10241328 | Germany | – | |
| 10328345 | Germany | A | |
| 10328345 | Germany | A | |
| 10328345 | Germany | – | |
| 0309814 | European Patent Office (EPO) | W | |
| 0309814 | European Patent Office (EPO) | W | |
| 10241328 | – | – | – |
| 10328345 | – | – | – |
| DE2002141328 | – | – | – |
| DE2003128345 | – | – | – |
| EP2003009814 | – | – | – |
| WO2003EP09814 | – | – | – |
Members7
| Document | Office | Kind | |
|---|---|---|---|
| DE10241328A1 | Germany | A1 | |
| WO2004023382A1 | World Intellectual Property Organization (WIPO) | A1 | |
| AU2003273397A1 | Australia | A1 | |
| EP1535237A1This record | European Patent Office (EPO) | A1 | |
| DE10241328B4 | Germany | B4 | |
| US2006161992A1 | United States of America | A1 | |
| US7729520B2 | United States of America | B2 |
8 legal events, as the office reported them to INPADOC
Over the term
Point at a mark for the eventEvents
| Event | Code | |
|---|---|---|
| Application deemed to be withdrawnWithdrawn18D | 18D | |
| Information on the status of an ep patent application or granted ep patentGrantedSTATUS: THE APPLICATION IS DEEMED TO BE WITHDRAWNSTAA | STAA | |
| First examination report despatched17Q | 17Q | |
| Request for extension of the european patent (deleted)DAX | DAX | |
| Request for examination filed17P | 17P | |
| Designated contracting statesAK | AK | |
| Request for extension of the european patentAX | AX | |
| Public reference made under article 153(3) epc to a published international application that has entered the european phaseORIGINAL CODE: 0009012PUAI | PUAI |
Numbers
- Publication
- 1535237
- Publication, DOCDB
- 1535237
- Publication, EPODOC
- EP1535237
- Application
- 3755552
- Application, DOCDB
- 03755552
- Application, EPODOC
- EP20030755552
Titles3
- German
- BIOMETRISCHES AKUSTISCHES SCHREIBSYSTEM UND VERFAHREN ZUR PERSONENIDENTIFIKATION UND HANDSCHRIFTENERKENNUNG MITTELS BIOMETRISCHER DATEN
- English
- BIOMETRISCHES AKUSTISCHES SCHREIBSYSTEM UND VERFAHREN ZUR PERSONENIDENTIFIKATION UND HANDSCHRIFTENERKENNUNG MITTELS BIOMETRISCHER DATEN
- French
- SYSTEME D'ECRITURE ACOUSTIQUE BIOMETRIQUE ET PROCEDE D'IDENTIFICATION DE PERSONNES ET RECONNAISSANCE DE L'ECRITURE A L'AIDE DE DONNEES BIOMETRIQUES
Classification
- CPC, 1
- G06V40/30
- IPC, 1
- G06K9 00
Designated states2
- Contracting states, 1
- Türkiye
- Extension states, 1
- North Macedonia