Advances in Electrical and Computer Engineering | |
Visual Peoplemeter: A Vision-based Television Audience Measurement System | |
SKELIN, A. K. ; SUPUK, T. G. ; BONKOVIC, M.. | |
关键词: consumer behavior; machine vision; object detection; image motion analysis; video surveillance; | |
DOI : 10.4316/AECE.2014.04011 | |
学科分类:计算机科学(综合) | |
来源: Universitatea "Stefan cel Mare" din Suceava | |
【 摘 要 】
Visual peoplemeter is a vision-based measurement system that objectively evaluates the attentive behavior for TV audience rating, thus offering solution to some of drawbacks of current manual logging peoplemeters. In this paper, some limitations of current audience measurement system are reviewed and a novel vision-based system aiming at passive metering of viewers is prototyped. The system uses camera mounted on a television as a sensing modality and applies advanced computer vision algorithms to detect and track a person, and to recognize attentional states. Feasibility of the system is evaluated on a secondary dataset. The results show that the proposed system can analyze viewer's attentive behavior, therefore enabling passive estimates of relevant audience measurement categories.
【 授权许可】
Unknown
【 预 览 】
Files | Size | Format | View |
---|---|---|---|
RO201902180500323ZK.pdf | 1KB | download |