视频中人像静态特征和动态特征研究-海洋环境科学.PDF

刑事技术 2015 年 第40 卷 第3 期 ·论  著· DOI ：10.16467/j.1008-3650.2015.03.012 视频中人像静态特征和动态特征研究 白笙学，刘 涛，王 昕，杨 鸣 （重庆市公安局物证鉴定中心，重庆 400021） Analysis of Human Static and Dynamic Characteristics in Video BAI Shengxue, LIU Tao, WANG Xin, YANG Ming Institute of Forensic Science, Chongqing Public Security Bureau, Chongq - ( ing , China 400021 ) ABSTRACT: Video investigation focuses on the videos involved in criminal cases and the relevant clues could be obtained through analyzing the persons and objects in videos. A video is composed of multi frame images, and characterized by intuitiveness and dynamic continuity. As the video provides both the static and dynamic human feature, some important characteristics can be extracted from the video, which are favorable for criminal investigation. In this paper, it is expounded that the main content of video image analysis and inspection of a human is to extract and analyze the static and dynamic characteristics. With video of high quality, the static characteristics such as the appearance and posture become the key points of video image analysis and inspection. In reality, the surveillance video is often apt to be influenced by various factors, such as lamplight, bad weather, distance, image distortion. The videos recorded by the monitoring system often tend to be blurring and it is difficult to extract the suspect’s facial features from those poor quality videos. Therefore, it is unlikely to get enough information if only relying on the static characteristics. However, movement and gait of the suspect could be totally recorded by surveillance video, which are less susceptible to the external environment. It is advantageous to combine human static and dynamic chara