Abstract
In this paper, we propose an intelligent pedestrian-search system for the video captured by a CCTV camera at a long distance. Our system detects and examines the target-pedestrians in real-time according to the query of their physical attributes. The main contributes of our system are summarized as: (i) stable tracking of the pedestrians from the long-distance images, (ii) fast extraction of the attribute features of pedestrians, and (iii) real-time detection of the pedestrians who wear the clothes that satisfy given color queries. Our system reliably searches and tracks the target-pedestrians in realtime.
| Original language | English |
|---|---|
| Pages (from-to) | 2193-2206 |
| Number of pages | 14 |
| Journal | Information |
| Volume | 15 |
| Issue number | 5 |
| State | Published - May 2012 |
Keywords
- Color classification
- Intelligent surveillance system
- Object tracking