Reconstruction attack method for biological template protection based on generative adversarial network

Abstract

The invention discloses a reconstruction attack method for biological template protection based on a generative adversarial network. The method comprises the steps: 1) basic data processing which comprises image cutting, feature extraction, hash code generation and training set and test set division; 2) constructing a biological hash reconstruction attack network which comprises a hash network anda generative adversarial network, the generative adversarial network comprising two stages, the first stage generating a low-resolution image through a given hash code, and the second stage generating a high-resolution image in combination with the low-resolution image and the given hash code; 3) updating the hash network weight; according to the method, semantic preservation and visual similarity are combined, the biological image is reconstructed from the given hash code through the hash network and the generative adversarial network, verification can be conducted through a biological recognition system, and high visual similarity can be kept between the biological image and a real biological image.

Description

technical field [0001] The invention relates to the technical field of biometric identification, in particular to a reconstruction attack method for biometric template protection based on a generative adversarial network. Background technique [0002] With the wide application of biometric systems including iris, face, fingerprint, voice, etc. a field. In order to avoid the irreparable loss caused by the leakage of biological template information, researchers have proposed many different biological template protection schemes. The difference between the biometric system based on template protection and the original biometric system lies in the feature extraction unit and the matcher. A template protection unit has been added. Generally speaking, an ideal biological template protection scheme needs to meet the following four characteristics: 1. Irreversibility: it is difficult to infer the original template information from the protected template information; 2. Revocabilit...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art, and propose a reconstruction attack method based on generative confrontation network for biological template protection, so as to solve the problem that the reconstructed image cannot meet the visual authenticity in the traditional reconstruction attack method , the reconstructed real biological image using a given hash code can not only pass the verification of the biometric identification system but also maintain a high visual similarity with the real biological image. At the same time, it inspires researchers to introduce necessary mechanisms to prevent possible security risks in the design process of related biological hash algorithms.

Images