Enhanced people detection combining appearance and motion information

The combination of two of the most recent people detectors from the state of the art is proposed. It is already known that the combination of independent information sources is useful for any detection task. In relation to people detection, there are two main discriminative information sources that characterise a person: appearance and motion. Proposed is the combination of two recent approaches based on both information sources. Experimental results over an extensive dataset show that the proposed combination significantly improves the results.

[1]  Bernt Schiele,et al.  Pedestrian detection in crowded scenes , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).

[2]  Álvaro García-Martín,et al.  People detection based on appearance and motion models , 2011, 2011 8th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS).

[3]  Álvaro García-Martín,et al.  On collaborative people detection and tracking in complex scenarios , 2012, Image Vis. Comput..

[4]  David A. McAllester,et al.  Object Detection with Discriminatively Trained Part Based Models , 2010, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[5]  Bill Triggs,et al.  Histograms of oriented gradients for human detection , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).

[6]  Álvaro García-Martín,et al.  A corpus for benchmarking of people detection algorithms , 2012, Pattern Recognit. Lett..