Study On Visual Comfort Evaluation Method For 3D Video

With the improvement of science and technology,videos and images from no sound to sound,from black and white to color,from smooth to high definition,are rapidly changing step by step.The original two-dimensional images no longer meet the needs of people.Digital 3D video technology is coming gradually into lives of people and is widely used.3D video is also known as stereoscopic video.Compared with normal 2D video,the most important visual feature is that the video is no longer limited to the screen,but can display three-dimensional video from top to bottom,left and right,front and back.It gives viewers an immersive feel.At present,the rapid development of 3D technology has been gradually applied to many fields,such as three-dimensional video conference system,remote industrial control,virtual reality,education,military,medical treatment,automatic navigation and consumer electronics.3D technology brings a visual revolution to 3D experience.It brings people a three-dimensional visual experience while exposing various problems.For example,it is easy to cause dizziness and other visual discomfort when people watch 3D video for a long time.In the future development,whether 3D technology can be accepted by more people,providing good 3D effects is only one aspect,and more importantly,it should pay attention to the quality of experience of stereoscopic contents.One of the most concerned is the degree of visual comfort for 3D video.Under the current trend of rapid development of 3D technology,the problem of visual discomfort caused by 3D video has been paid more and more attention by researchers in related fields.Improving the comfort of 3D video has become a compulsory course,and making a corresponding visual comfort evaluation for 3D video has become a very important research topic.From the perspective of evaluation method,the evaluation of 3D video comfort can be divided into subjective evaluation and objective evaluation.At home and abroad related research reports,the stereoscopic video visual comfort evaluation method does not exist a unified standard.In the current background of the development of 3D video technology,this paper draws on the existing references and research results at home and abroad.Two objective visual comfort evaluation methods for 3D video are proposed through the analysis of the content of 3D video.The details are as follows:(1)A 3D video visual comfort evaluation method on the consistency of accommodation and convergence: First,the comfort evaluation factor based on the feature points of images is defined according to the horizontal disparities of each frame of left and right view for 3D video and the binocular comfortable disparity range determined by the viewing distance.Then two overall objective evaluation methods onvisual comfort for each frame and for multi frames on the consistency of accommodation and convergence are established according to the comfort evaluation factor.The innovation is that the proposed objective evaluation method on visual comfort for 3D video considers the binocular comfortable disparity range determined by the viewing distance and discusses the effect of different horizontal disparities and viewing distances on visual comfort for 3D video.(2)A 3D video visual comfort evaluation method based on disparity information and motion: Firstly,the horizontal and vertical disparity of each frame of 3D video and the motion vector of the adjacent frame are calculated.Secondly,horizontal angular disparity,vertical angular disparity and motion are calculated in angular units.Then,three objective evaluation models of horizontal angular disparity,vertical angular disparity and motion are established,respectively.Finally,two overall visual comfort measures are proposed for stereoscopic image and video,respectively.The innovation lies in that the proposed objective evaluation method calculates horizontal disparity,vertical disparity and motion in angular units.The calculation of angular can change with the viewing distance,display size and image size.Therefore,the established objective evaluation method and subjective evaluation results have a better correlation under different viewing conditions(viewing distance,display size and image size).The visual comfort evaluation model for motion is established considering three different situations,which are planar motion,depth motion and composite motion.We conduct different subjective evaluation experiments to verify the performance of two objective evaluation methods.The experimental results show that the two proposed objective evaluation models have good performance and good consistency with the subjective evaluation results.