In recent years, 3D technology has become an emerging industry. However, visual fatigue always impedes the
development of 3D technology. In this paper we propose some factors affecting human perception of depth as new
quality metrics. These factors are from three aspects of 3D video--spatial characteristics, temporal characteristics and
scene movement characteristics. They play important roles for the viewer's visual perception. If there are many objects
with a certain velocity and the scene changes fast, viewers will feel uncomfortable. In this paper, we propose a new
algorithm to calculate the weight values of these factors and analyses their effect on visual fatigue.MSE (Mean Square
Error) of different blocks is taken into consideration from the frame and inter-frame for 3D stereoscopic videos. The
depth frame is divided into a number of blocks. There are overlapped and sharing pixels (at half of the block) in the
horizontal and vertical direction. Ignoring edge information of objects in the image can be avoided. Then the distribution
of all these data is indicated by kurtosis with regard of regions which human eye may mainly gaze at. Weight values can
be gotten by the normalized kurtosis. When the method is used for individual depth, spatial variation can be achieved.
When we use it in different frames between current and previous one, we can get temporal variation and scene
movement variation. Three factors above are linearly combined, so we can get objective assessment value of 3D videos
directly. The coefficients of three factors can be estimated based on the liner regression. At last, the experimental results
show that the proposed method exhibits high correlation with subjective quality assessment results.