A Noise Resistant Method by Setting Thresholds for Use With Face Recognition

Face recognition is one of the important research contents of biometric recognition, and it is widely used in various fields of life. The recognized face image is acquired by devices such as video capture, and noise is generated by device factors or external influences during the acquisition process, thereby reducing the accuracy of recognition. If the noise intensity is basically fixed, the smaller the pixel value is, the larger the proportion of noise is. This pixel has low reliability for image description. Therefore, a method of noise resistant is proposed from the angle of image preprocessing to reduce the impact of this problem. This anti-noise method first sets a certain threshold. If the pixel value in the image is lower than it, the pixel values below the set value will be directly set to 0. Then, using the classification algorithm of cosine similarity, the similarity calculation is performed on the face image, so that the image is classified, finally recognized, and the results are output. The experimental results show that after the appropriate threshold is selected, the proposed method can reduce the interference of noise on the classification and improve the accuracy of face recognition.