Face recognition attack defense method based on Rosenbrock-PSO

Abstract

The invention discloses a face recognition attack defense method based on Rosenbrock-PSO. The method comprises the following steps: (1) collecting and preprocessing face image data, and dividing the face image data into a pre-training data set and a disturbance data set; (2) manufacturing a pure-color glasses frame template for limiting a disturbance area on the face image in the disturbance dataset; (3) training a face classifier of the face recognition model by using the pre-training data set; (4) constructing a Rosenbrock-PSO face attack model; (5) inputting a face image with a pure-colorglasses frame into the Rosenbrock-PSO face attack model for evolutionary optimization, to obtain a face image when RGB values on a glasses frame are optimal solutions, and the face image is used as anadversarial sample; and (6) adding the confrontation sample into the pre-training data set to retrain the face classifier, so that the face recognition model has the capability of defending against the attack of the confrontation sample. By utilizing the method and the device, the face recognition model which can defend countermeasure samples and has relatively high generalization ability can beobtained.

Description

technical field [0001] The invention belongs to the fields of computer vision and deep learning, and in particular relates to a face recognition attack defense method based on Rosenbrock-PSO. Background technique [0002] Face recognition is mainly to automatically extract face features from face images, and then perform identity verification based on these features. With the rapid development of machine learning, face recognition technology has been continuously improved, and the recognition accuracy has also been continuously improved, especially in business, it has been widely used, such as face payment, face attendance, face check-in, face opening, etc. . At the same time, face recognition technology can also be widely used in smart police and smart city construction, providing smart face services for the whole society. [0003] Although advanced face recognition technology provides great convenience to our daily life, recent studies have shown that advanced convolutio...

AI Technical Summary

Problems solved by technology

[0006] Aiming at the security problem that the current face recognition technology is vulnerable to adversarial attacks, the present invention provides a face recognition attack defense method based on Rosenbrock-PSO, which can obtain a face that can defend against adversarial samples and has strong generalization ability. recognition model

Images