Applications of Computer Vision, IEEE Workshop on (2013)
Clearwater Beach, FL, USA USA
Jan. 15, 2013 to Jan. 17, 2013
Rogerio Feris , IBM T. J. Watson Research Center, New York
Ankur Datta , IBM T. J. Watson Research Center, New York
Sharath Pankanti , IBM T. J. Watson Research Center, New York
Ming-Ting Sun , University of Washington, Seattle
We present a novel approach to automatically create efficient and accurate object detectors tailored to work well on specific video surveillance cameras (specific-domain detectors), using samples acquired with the help of a more expensive, general-domain detector (trained using images from multiple cameras). Our method requires no manual labels from the target domain. We automatically collect training data using tracking over short periods of time from high-confidence samples selected by the general-domain detector. In this context, a novel confidence measure is proposed for detectors based on a cascade of classifiers, which are frequently adopted for computer vision applications that require real-time processing. We demonstrate our proposed approach on the problem of vehicle detection in crowded surveillance videos, showing that an automatically generated detector significantly outperforms the original general-domain detector with much less feature computations.
Detectors, Feature extraction, Videos, Target tracking, Surveillance, Training, Cameras
R. Feris, A. Datta, S. Pankanti and M. Sun, "Boosting object detection performance in crowded surveillance videos," Applications of Computer Vision, IEEE Workshop on(WACV), Clearwater Beach, FL, USA USA, 2013, pp. 427-432.