OCT fingerprint section image authenticity detection method based on reconstruction difference

Abstract

The invention discloses an OCT fingerprint section image authenticity detection method based on reconstruction difference, and the method comprises the steps: S1, constructing a full convolutional neural network model which comprises an encoder, a generator and a feature extractor; s2, collecting images collected by an OCT system, and after preprocessing is completed, randomly selecting 70% of positive sample images as training data; selecting the other 30% of positive sample images and negative sample images, and taking the images as test data after quantity equalization; s3, training a network model; selecting the divided training image as input data, and setting a loss function for optimizing an encoder and a generator; setting comparison loss for optimizing the feature extractor; performing multi-round training on the established network model, updating and optimizing model weight parameters through back propagation until a loss function tends to converge, and stopping training; s4, testing the network model; and applying the trained network model, selecting a test data input model for testing, and performing authenticity judgment on an input image according to a set threshold value.

Description

technical field [0001] The invention relates to the technical field of biometric identification and abnormality detection, and is particularly applied to detection of forged OCT fingerprint slice images. Background technique [0002] A major feature of optical coherence tomography (OCT) is that it can detect two-dimensional or three-dimensional structural images of biological tissues. When applied to the finger, it can detect the subcutaneous information of the finger, which can not only be used to reconstruct the fingerprint and identify it, but also improve the ability of living detection, and has a certain degree of anti-counterfeiting ability. However, the current fingerprint identification system based on OCT usually needs manual participation to judge the authenticity of the image after collecting the image, and still lacks an efficient and accurate automatic identification method. [0003] Forgery sample detection is a specific application in anomaly detection. In r...

AI Technical Summary

Problems solved by technology

[0004] The present invention overcomes the disadvantages of the prior art in identifying counterfeit OCT fingerprint slice images, and provides a method for detecting counterfeit fingerprint OCT slice images that is simple, automatic, and does not require a large amount of complicated preprocessing

Images