基于低秩矩阵分解的视频前景目标提取问题研究
Research on the Video Prospect Objective Extraction Problem Based on Low-Rank Matrix Factorization

监控视频在安防领域和视频内容理解中具有重要作用，而准确提取监控视频中的前景目标对于实现有效的视频内容理解、事件检测和场景分析至关重要。然而，监控视频通常面临着复杂、动态的背景干扰以及目标的多变性，这给前景目标提取带来了一定的挑战。针对这一问题，本文总结了3种监控视频前景目标的提取与处理方法，并对其进行了研究与实现。
Surveillance video plays an important role in the field of security and video content understanding, and the accurate extraction of the prospect goals in surveillance video is crucial to achieving effective video content understanding, event detection and scene analysis. However, surveillance video is often faced with complex and dynamic background interference as well as variability of targets, which brings some challenges for foreground target extraction. To solve this problem, this paper summarizes the extraction and processing methods of three surveillance video prospect targets, and studies and realizes them.