Intelligent dermatoglyph collecting, classifying and recognizing method based on neural network

Abstract

The invention relates to an intelligent dermatoglyph collecting, classifying and recognizing method based on a neural network. The method comprises the following steps: sequentially carrying out normalization processing and Wiener filtering denoising processing on an acquired original dermatoglyph image, sharpening and extracting edges by adopting a Sobel operator algorithm, carrying out processing by adopting a binarization algorithm, and obtaining a pixel-skeletonized dermatoglyph image through an OPTA refinement algorithm; then, adopting a GAN generative adversarial network model algorithmto carry out repair processing and enhancement processing; and finally, introducing a ResNet deep learning neural network model algorithm to carry out dermatoglyph classification and recognition processing, and obtaining classification information. According to the method, different dermatoglyph feature images are analyzed from the perspective of multiple dimensions and multiple features, more features are extracted from dermatoglyph image information, and high accuracy is achieved in dermatoglyph recognition and classification; the method is suitable for being applied to various scenes such as child personality culture and special post personnel selection, the interpretation accuracy is improved, and the reference of report interpretation is obviously improved.

Description

technical field [0001] The invention relates to the classification and identification technology of dermatoglyphs, in particular to an intelligent dermatoglyphs collection, classification and identification method based on neural network. Background technique [0002] Skin lines refer to raised lines on fingers, palms, toes, and soles. Our common name "fingerprint" refers to the lines on fingers, and the lines on palms and soles are called palm lines and foot lines, respectively. Once the dermatoglyphs are developed, they will remain unchanged until now. The most effective feature that most of the existing fields rely on when applying dermatoglyphics technology is its uniqueness. It can be widely used in criminal investigation, medicine, biology, etc. Identify fields. [0003] First of all, its uniqueness is reflected in its uniqueness that distinguishes it from other skin patterns. Not only is it different from others, but also its ten fingerprints are also different, incl...

AI Technical Summary

Problems solved by technology

[0013] The first factor is the part of pattern entry. The current applications of various dermatoglyph technologies are aimed at adults, and the collector is also designed for adults. Since the fingers of young people are generally thinner, the development of dermatoglyphs is not easy. Mature, features are not obvious, traditional fingerprint collectors can not meet the requirements of collection and identification, young children are uncontrollable, if a large number of batch collection work is carried out, it cannot be realized smoothly;

[0014] The second factor is the skin texture collection process. Since the shooting environment has a great influence on the image effect, such as lighting, angle, and pixels, the collected skin texture is likely to be damaged, which may lead to repeated collection. Or the effect of recognition and classification is not ideal due to the unclear and incomplete collected images, and there is a lack of introduction of corresponding program algorithms for repair;

[0015] The third factor is that there is a large amount of grain types, which can be divided into more than 20 types, and the differences between grain types are relatively small. Using a manual method requires a lot of complicated professional learning and a long period of time. Practice can master the ability of pattern recognition

Obviously, the manual method will lead to inaccurate recognition of the pattern, the final report and the analysis and interpretation of the report are not accurate enough, and the report conclusion is less referable;

[0016] The fourth factor is that fingerprint recognition classification generally uses a shallow structure model to process data, and the structure model has at most one or two layers of nonlinear features. The shallow structure model has been used to solve some simple practical problems, but When encountering complex multi-dimensional and multi-feature situations, it is difficult for shallow models to express well;

[0017] The fifth problem is the complexity of report interpretation. Due to irregular collection, classification and identification, the final report is huge in content, with many parameters and indicators, making it difficult to conduct comprehensive analysis. Experts are required to do professional one-on-one Interpretation

Images