Automatic 3D video detection

In this paper, we propose a novel and effective method to detect whether a video is 3D Side-by-Side format or not. The proposal extracts features between left and right side of video frames for detection. We firstly obtained the difference between left and right side of video frame and adopted a threshold operation for denoising. Secondly, we introduced a ratio method to improve the low precision. Finally, some frames which meet certain conditions were chosen for decision. The experimental results indicate that our proposed method can perform well. The best performance we obtained is 2.88% for false negative rate as 3.40% for false positive rate, by using 1,800 videos for evaluation.