A method for detecting and locating standing faces in an intelligent recording and broadcasting system

Abstract

The invention belongs to the field of intelligent recording and broadcasting systems, and relates to a method for detecting and locating standing human faces in an intelligent recording and broadcasting system. value, and then perform morphological operations on the image, mark the active area and retain it with the smallest circumscribed rectangle, and perform light compensation on it, and then filter out the interfering active area through skin color detection, and classify it according to the proportion of the face area Screening and exclusion of skin-colored but no-face areas, and finally data analysis based on the processed data to mark faces. This method is ingenious in design, simple in detection principle, easy to operate and use, and can complete standing with only one camera. Face detection and positioning, low equipment requirements, high accuracy of detection results, friendly application environment, and broad market prospects.

Description

Technical field: [0001] The invention belongs to the field of intelligent recording and broadcasting systems, and relates to a method for detecting and locating human faces in a recording and broadcasting system, in particular to a method for detecting and locating standing human faces in an intelligent recording and broadcasting system, which combines illumination compensation on the basis of frame difference method , skin color detection, and morphological processing methods to determine the standing position of students through face detection. Background technique: [0002] With the development and advancement of education informatization and quality course construction, intelligent recording and broadcasting systems are widely used in classroom teaching and teaching research. As the main body of classroom teaching activities, students' classroom performance and the entire teaching process guided by teachers are not only the focus of teaching activities, but also the focu...

AI Technical Summary

Problems solved by technology

Zhang Shu et al. extracted the local area of ​​the face as a training sample, learned a dictionary with strong discriminative properties, judged whether a certain local area of ​​the face appeared based on the response of the sparse coding of each detection window, and used the detection results and position of the local area of ​​the face Constraint voting to complete face positioning; Xu Yan et al. used the method based on boundary tracking and multi-angle imaging to position the camera, combined with the DWT algorithm for face detection and feature positioning; Wang Xiaoyu et al. proposed a method based on multi-block local binary Adaboost algorithm based on value pattern features and human eye positioning method based on template matching; Yu Jialin proposed an image stabilization algorithm based on motion imaging model and adaptive mean motion filter algorithm for the detection of video face images; Liu Wangsheng et al. proposed an improved AdaBoost face detection algorithm, the algorithm comprehensively uses human skin color model, face motion detection model and improved background extraction method to detect faces; this algorithm improves the detection speed, but the robustness to light mutation is not high, Ye Xueyi et al. proposed a A detection method based on a probabilistic multi-layer restricted Boltzmann machine (RBM) cascaded neural network. This method takes P-RBM as the core, forms a multi-hidden layer neural network by cascading multiple P-RBMs, and then combines hidden The state of each neuron in the layer realizes the face detection under the non-ideal state, but none of the above methods can solve the problem of how to quickly detect and locate the standing person in the squatting crowd. In the prior art, the application number is CN201110032341. The Chinese patent of 3 discloses a method for detecting students sitting up in a teaching intelligent recording and broadcasting system. At least two cameras are set in the area where students stand up. The detection steps are: image acquisition and calibration. Each video camera is all demarcated student's stand-up area; Motion detection, has adopted moving part to detect, and uses the subtraction between images; Student detection, the advantage of the present invention is that does not need to refit student seat, is easy to maintain, but the present invention needs many Simultaneous operation of two cameras, the equipment is relatively complex, and the working environment requires high requirements. Therefore, a method of standing face detection and positioning in the intelligent recording and broadcasting system is designed and prepared. The face detection and positioning can be completed through the camera, and the practical performance is good.

Images