Multi-scale visualization based on sketch interaction for massive surveillance video data

Zhengming Zhang, Ran Zuo, Rui Guo, Yanfeng Li, Ti Zhou, Han Xue, Cuixia Ma, Hongan Wang

Abstract: Large amount of surveillance video data needs effective and interactive methods to record and utilize. Traditional methods mainly focus on low-level analysis of surveillance video data and lack user interaction. Therefore, there is a strong demand of easy-to-use interaction for efficiently analyzing surveillance video information. In this paper, we propose a multi-scale approach merged on the data flow, objects, and frames to achieve visualization of surveillance video data. Combined with the advantage of sketch interaction, the design of multi-scale structure makes the analysis of surveillance content natural and fluent with annotation of video contents. Extensive user studies demonstrate the effectiveness for facilitating users’ interactive visual analysis in surveillance video exploration and significantly reduce playback time to confirm available information.