Research On Image Quality Detection For Video Conference

With the rapid development of network technology and multimedia technology, people broke through the traditional conference mode by using video conference. Video conference system has been an important communication services, which is used by all walks of life including policy promotion, management, emergency command and training. With the continuous development and expansion of video conference systems, the quality of video conference gradually aroused people's attention. Video image quality distortion has a negative effect on meetings, which can even lead an interruption to the video conference. In this paper, we study the six types of video distortion, including noise,blur,color cast,abnormal brightness, screen freeze (static frame) and black screen. Then we present an image quality detection system for video conference.The detection methods of abnormal brightness, screen freeze (still frame) and black screen are relatively simple. So we focus on the detection methods of noise, blur and color cast. Then we propose new detection methods for these three distortions respectively. In the aspect of noise detection, several methods of image noise detection which have been applied in practice are introduced. After considering the characteristics of video conference, a simple and efficient noise detection method for video conference—an image noise detection method based on 2D entropy and mean square error of block is presented. In the aspect of blur detection, the existing methods of image blur evaluation are discussed. After comparison, we decide to use the image blur evaluation method based on structural similarity index measurement of gradient image (GSSIM) to detect the video of conf-erence. In order to improve the adaptability of GSSIM, an image blur detection method based on multi-scale GSSIM is proposed. In the aspect of color cast detection, we propose an image color cast detection method based on point sharpness of CIELab to overcome the shortcomings of the existing image color detection method. Experiment results show that the above three methods can detect the conference video in real time and obtain the relatively accurate test result.Finally, the above three kinds of detection methods are applied to video conference system. An image quality detection scheme which can detect the common six types of distortion for video conference system is presented. We use C ++ to write code to implement the system design on VS2013 with OpenCV.