Surveillance video fast browsing

• Main Author: Wei, Hao
• Other Authors: Wang Shizheng
• Format: Final Year Project
• Language: English
• Published: 2015
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/65801
• Institution: Nanyang Technological University

In this paper, I will be introducing a video analysis technique namely Video Synopsis as well as a Privacy Preserving Framework for Surveillance Video Browsing. Basically, video synopsis helped to a great extent reducing the workload for a video reviewer to examine a particular video stream for useful information which is especially time-consuming. Video synopsis technique is able to transform a long cumbersome surveillance video into an abstracted essential video with all information from the detailed-based video analysis which generates and stores a set of object flags for further browsing use. Compare to the conventional surveillance video browsing, we have improved the video reviewing practice mainly in three ways: valuable scene highlighting, higher storage efficiency and multi-scale scalable browsing. Secondly, on the basis of video synopsis technique, we have purposed a privacy preserving framework on the moving object segmentation in the compressed domain for ease of ensuring the owner’s or service provider’s privacy information. Extensive experiments on H.264 standard bitstreams demonstrate considerable performance improvements in privacy preserving in the area of intelligent video surveillance.