Fingerprint identification model construction method, storage medium and computer equipment

Abstract

The application provides a method for constructing a fingerprint recognition model based on Resnet and Triplet Loss, which includes constructing N groups of initial samples; using Triplet Loss to train N groups of initial samples to obtain N groups of training samples; inputting N groups of training samples into the initial In the Resnet model, the initial Resnet model is trained to obtain the target Resnet model to be tested; two sets of test samples are input into the Resnet model of the target to be tested for calculation to obtain two sets of image feature quantities; the preset algorithm is used to calculate the difference between the two sets of image feature quantities The maximum cosine similarity between; determine whether the maximum cosine similarity is greater than or equal to the preset value; generate the target Resnet model according to the target Resnet model to be tested; or construct M groups of training samples and input the M groups of training samples to the target to be tested The target Resnet model to be tested is trained in the target Resnet model. The application also provides a storage medium and computer equipment. The above method effectively solves the problem of narrow-edge fingerprint matching and recognition through the fingerprint recognition model jointly constructed by Resnet and Triplet Loss.

Description

technical field [0001] The present application relates to the field of fingerprint identification, in particular to a method for constructing a fingerprint identification model, storage media and computer equipment. Background technique [0002] In today's society, with the continuous and rapid development of science and technology, mobile phones, computers and other electronic equipment are frequently updated, and the use of fingerprints for identity verification and identification has become popular. Fingerprint is a highly recognizable and unique human biological feature, and has the characteristics of easy collection and high security. Fingerprint identification is one of the main methods of current biometric identification technology. With the development of biometric technology, the fingerprint recognition technology under the screen has a tendency to change to the side fingerprint. Side fingerprint collection requires devices with smaller sensing area and more conven...

AI Technical Summary

Problems solved by technology

Side fingerprint collection requires devices with smaller sensing area and more convenient application. However, the collected fingerprint images are slender and strip-shaped, and there is little useful fingerprint information. The direction and position of entering and unlocking fingerprints may vary greatly. Bringing challenges to rejecting authenticity and identifying falsehoods in fingerprint identification

Images