An Automatic Signature Discrimination Method Based on Fuzzy Mean Hash Learning

Abstract

The invention discloses an automatic signature discrimination method based on fuzzy mean value hash learning, which relates to the field of computer application technology, including step 1, preprocessing the images in the signature positive data set, learning image features, and using the convolutional neural network algorithm to Image features are extracted from images in the signed positive dataset. The application device of the present invention is diversified, that is, the server system can be directly accessed through the mobile terminal, and the server system can also be connected through the electronic screen, which is applicable to a wide range of applications. The present invention can automatically determine whether the signature is the person's own, which improves the accuracy of identification The invention automatically learns the user's signature model, simplifies the difficulty of distinguishing authenticity, reduces the burden of experts, and performs hash learning through real-time processing, which greatly reduces the overhead of computer memory and reduces the hardware pressure on the server.

Description

technical field [0001] The invention relates to the field of computer application technology, in particular to an automatic signature discrimination method based on fuzzy mean value hash learning. Background technique [0002] At present, when the staff compares whether the signature is the user's own signature at work, there are problems such as difficult identification process, insufficient user privacy security, and lack of methods for comparing the degree of discrimination of similar images. Moreover, the general signature identification system requires a lot of experts. Resources, the training cost of experts in this field is very high, and they are relatively scarce in life. [0003] Most of the existing image systems use approximate neighbor search, which lacks the analysis of specific areas of difference in similar images; most of the existing hash learning methods are used in the field of approximate neighbor retrieval, and there is no method for distinguishing simi...

AI Technical Summary

Problems solved by technology

[0002] At present, when the staff compares whether the signature is the user's own signature at work, there are problems such as difficult identification process, insufficient user privacy security, and lack of methods for comparing the degree of discrimination of similar images. Moreover, the general signature identification system requires a lot of experts. Resources, the training cost of experts in this field is very high, and they are relatively scarce in life

[0003] Most of the existing image systems use approximate neighbor search, which lacks the analysis of specific areas of difference in similar images; most of the existing hash learning methods are used in the field of approximate neighbor retrieval, and there is no method for distinguishing similar images, and the existing The hash method is difficult to deal with the online processing of data and lacks a feedback mechanism. Hash learning generally requires a large amount of training data, but it is very difficult to collect a large number of signature data of the same person in reality

Images