Real-time approximate frontal face image optimizing method and system based on several cameras

Abstract

The invention discloses a real-time approximate frontal face image optimizing method and system based on several cameras. The real-time approximate frontal face image optimizing system comprises a three-camera system capable of adjusting angle and a software system calculating frontal face direction confidence score by multiple threads. The software system detects a human face area and tracks the human face by the CT algorithm, and by adopting SDM algorithm matching characteristic points, confidence score data is given. The confidence score is valued within 0.0 to 1.0 and directly shows frontal direction degree of the human face, and the larger the value is, the closer the human face to the frontal direction. In the main thread, confidence scores of three threads are compared by the algorithm uninterruptedly, the images of the maximal confidence score are used output by the system, and the frontal face images are automatically optimized in real time. The method and the system can be used for information preprocessing of a face identifying and tracking system, and provides excellent frontal face image data for subsequent analysis processing.

Description

technical field [0001] The present invention relates to computer vision face image processing technology, in particular to a method and system for real-time automatic optimization of quasi-frontal face images based on multi-cameras, which can be used for information preprocessing of face recognition and tracking systems, and provide information for subsequent analysis and processing. Good quasi-frontal face image data. Background technique [0002] Face detection technology is one of the important branches in the field of computer vision. In recent years, it has been widely used in visual monitoring, video retrieval, robot control, human-computer interaction, etc., and has become a hot research topic in pattern recognition. The main purpose of face detection technology is to detect and locate a human face from an image, that is, to determine whether there is a human face in the image, and if so, to locate the human face area from the image. These located face area images ca...

AI Technical Summary

Problems solved by technology

[0009] Secondly, in the existing multi-camera face detection method, although there are multiple cameras facing the detected object, the probability of capturing a frontal image of a human face increases, and it is relatively easy to obtain a frontal image or a quasi-frontal image of a human face. However, the existing methods only focus on detecting and tracking the same face, and do not give quasi-frontal face images in real time, that is, they do not solve the real-time optimization problem of quasi-frontal faces. Therefore, in the face detection technology based on multi-camera In , multiple cameras will capture images of the same face in different directions at the same time, how to select the most accurate frontal face image is the key to the problem

Images