In this paper, a fast and low cost human identification procedure is proposed for an outdoor video-based surveillance. It consists of an effective background extraction and a skeleton based people detection with shape constrains. This procedure can output the number and the coordinates of those detected people for each frame of the video. The output can either be used by a data analysis system or a crowed modelling system. Satisfactory results obtained from a benchmark video stream had illustrated the effectiveness of the proposed approach.
|Title of host publication||RPC 2010 - 1st Russia and Pacific Conference on Computer Technology and Applications 2010|
|Number of pages||5|
|Publication status||Published - 01 Dec 2010|
|Event||1st Russia and Pacific Conference on Computer Technology and Applications, RPC 2010 - Vladivostok, Russian Federation|
Duration: 06 Sep 2010 → 09 Sep 2010
|Conference||1st Russia and Pacific Conference on Computer Technology and Applications, RPC 2010|
|Period||06/09/10 → 09/09/10|