Fast Pedestrian Detection Using Color Information
In a pedestrian detection system, the application of color information can increase the detection rate; however, the detection speed will be slowed down a lot. This paper presents a fast pedestrian detection method using color information. It firstly scans a pair of sequential gray-scale frames to select candidates using both appearance and motion features; and then uses information of each color channel (RGB) to do a further confirmation with support vector machine based classifiers. Compared with pedestrian detection systems that only use gray-scale information, the system using our method has almost the same detection speed; at the same time, it also gets a better detection rate and false-positive rate. The experiment in a pedestrian detection system with a single optical camera proves the effectiveness of our method.
KeywordsSupport Vector Machine False Positive Rate Support Vector Machine Classifier Color Information Color Channel
Unable to display preview. Download preview PDF.
- 2.Viola, P., Jones, M., Snow, D.: Detecting Pedestrians Using Patterns of Motion and Appearance. In: International Conference on Computer Vision (ICCV), vol. 2, pp. 734–741 (2003)Google Scholar
- 3.Papageorgiou, C., Poggio, T.: Trainable pedestrian detection. In: International Conference on Image Processing, vol. 4, pp. 35–39 (1999)Google Scholar
- 4.Joachims, T.: Making large-scale SVM learning practical. In: Schölkopf, B., Burges, C.J.C., Smola, A.J. (eds.) Advances in Kernel Methods – Support Vector Learning. MIT Press, Cambridge (1998)Google Scholar