Method and system for identifying human face in video play

Abstract

The invention discloses a method and a system for identifying a human face in video play, wherein the method comprises the following steps: A. constructing a diversity face database based on time domain continuous transformation; B. improving the depth learning facial feature extraction network to train a depth learning model base on time domain variability; C. performing face recognition throughfeature extraction and model classification. The face recognition method in the video playing of the invention constructs a diversity sample library with continuous change of time sequence, adds a GDCConv depth neural network learning unit, and reduces the influence of face angle deflection, expression, illumination on face image feature extraction. Finally, the identification information can be further confirmed by joint feature matching and model classification, which can effectively improve the accuracy of video face recognition.

Description

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine

Login to view more

AI Technical Summary

Benefits of technology

This patented technology uses advanced techniques such as convolutional networks (CNNs) or generators to extract important facial characteristics from videos without affecting their quality significantly compared to previous methods that only analyzed them once. It also includes an algorithm called Deep Learning Convolution Network(DCN), which helps identify faces more accurately than previously possible due to its ability to learn complex patterns over long sequences of data. Overall, this new technique allows for efficient processing of realistically captured images at home or work environments.

Problems solved by technology

This patented technical problem addressed by the present inventors relates to improving image processing techniques used during facial recognition (FAC) processes that involve capturing high quality images with limited motion). Current FAC systems rely heavily upon frame or region based registration strategies, making it difficult to identify any significant differences caused by factors like lighting conditions, pose variations, shape, texture, and other variables affecting the appearance of faces captured through cameras.

Images