Under-sample face recognition method based on multi-dimensional scale transformation network and block weighting method

Abstract

The invention discloses an under-sample face recognition method based on a multi-dimensional scale transformation network and a block weighting method. The steps are as follows: firstly, each image in a sample set composed of a single face image is subjected to block processing to obtain A new sample data set; then, use the new sample data set to learn the filter parameters of the multi-dimensional scale transformation network, and use the filter parameters to extract the feature expression of the sample image, and build the corresponding feature library; then, by calling the filter parameters , feature extraction is performed on the divided test image data set, and the extracted features are synthesized in a weighted manner, and then matched with the features in the feature library; finally, the final test face is obtained by using the matching results Image classification identification information. The invention uses an unsupervised feature extraction network framework to accurately extract face image features, thereby improving the accuracy of face recognition and laying a solid foundation for the construction of public safety.

Description

technical field [0001] The invention relates to the technical field of deep learning applications, in particular to an under-sample face recognition method based on a multi-dimensional scale transformation network and a block weighting method. Background technique [0002] In recent years, video surveillance has been popularized in large and medium-sized cities across the country, and has been widely used in the construction of social security prevention and control systems, and has become a powerful technical means for public security organs to investigate and solve crimes. Especially in mass incidents, major cases and double robbery cases, evidence clues obtained from video surveillance play a key role in the rapid detection of cases. Due to the impact of shooting time, space and environment, the face images that can be captured are diverse. Facts have proved that using one face image sample per person to identify face images under different lighting conditions, different...

AI Technical Summary

Problems solved by technology

For the under-sampled face recognition problem, the captured face images often have factors such as different lighting and different expressions, and there are too few label data available for training, which seriously affects the accuracy of face recognition.

Images