Filtering method and device for adversarial sample image, electronic equipment and medium

Abstract

The invention relates to an adversarial sample image filtering method and device, electronic equipment and a medium, and the method comprises the steps: obtaining an original adversarial sample image which comprises interference information; performing feature extraction on the original adversarial sample image to obtain first feature information of at least two dimensions, and performing singular value decomposition on the first feature information of each dimension to obtain second feature information corresponding to the first feature information of the dimension; for the second feature information of each dimension, determining third feature information meeting a filtering condition in the second feature information according to the preset filtering condition corresponding to the dimension; and determining a target confrontation image according to the third feature information of each dimension. Through the method provided by the invention, interference information in the original adversarial sample image can be effectively filtered out.

Description

technical field [0001] The present invention relates to the technical field of computer vision, and in particular, the present invention relates to a method, device, electronic device and medium for filtering an image of an adversarial sample. Background technique [0002] In recent years, with the rapid development of computer image recognition, various industries are trying to integrate computer vision into social production, with a wide range of applications. It has developed many directions not limited to image recognition, target detection, semantic segmentation, entity segmentation, etc., and has been applied to many aspects such as autonomous driving, medical image analysis, and face recognition. With the gradual maturity of the computer vision application field, the security of neural network models has gradually become an important research direction. [0003] As a basic application of computer vision, image recognition plays a very important role. The rapid devel...

AI Technical Summary

Problems solved by technology

Therefore, in the prior art, there is an urgent need to provide a method for filtering the interference information in the adversarial sample image to solve the problem that the prior art cannot accurately filter out the interference information in the adversarial sample image, and the model trained based on the adversarial sample image less robust problem

Images