Signature recognition system and method

Method of identification board comprises the following steps: Samples to the seal and storing indicated the data of the board, converting the data higher-dimensional vector, which the non-supervision neural network the higher-dimensional vector, performing senior and device for extracting wire to a higher-dimensional the position of the higher-dimensional vector from the group; and analysis higher-dimensional and cluster, according to the information according beforehand storage authenticity of the conversion judging with board. The device for the device for identifying comprising: For samples to the seal and storing indicated the data sampling device according board, and downstream of the sampling device, wherein the data to conversion higher-dimensional vector the switching device, for receiving higher-dimensional vector, and higher-dimensional vector performing senior and device for extracting wire, the identifying higher-dimensional of cluster's non-supervision neural network, and connected with the non-supervision neural network, for analyzing the higher-dimensional and data of the authentic analysis device for judging signed.