Video saliency detection method

The invention belongs to the technical field of computer vision, and discloses a video saliency detection method, which comprises the following steps of: extracting the depth, bit allocation, motion vector and other characteristics of all videos from corresponding VVC bit streams by using a sight tracking database, and calculating. A map of these features is then generated for each video frame. Then, the features are optimized through a filter, and data are trained through a nonlinear dichotomy method; and finally, saliency detection is completed through saliency mapping and processing, so that saliency detection is rapidly and accurately carried out, the calculation amount of an algorithm is reduced, and the detection efficiency is improved. 本发明属于计算机视觉技术领域，公开了一种视频显著性检测方法，通过使用视线跟踪数据库，对其中的所有视频的深度、位分配和运动向量等特征从相应的VVC比特流中提取并进行计算。接着，为每个视频帧生成这些特征的地图。然后，将上述这些特征经过滤波器进行优化，并通过非线性二分类的方法训练数据。最后，通过显著性映射和处理，完成显著性检测，从而实现了快速准确地进行显著性检测，减小算法计算量，提高检测效率。