Title :
Cascading Rectangle and Edge Orientation Features for Fast Pedestrian Detection
Author :
Chen, Yu-Ting ; Chen, Chu-Song
Author_Institution :
Acad. Sinica, Taipei
Abstract :
In this paper we develop a pedestrian detection method that can detect human in a single image based on a boosted cascade structure. In our approach, both the rectangle features and 1-D edge-orientation features are employed in the feature pool for weak-learner selection, which can be computed via the integral-image and the integral-histogram techniques, respectively. To make the weak learner more discriminative, Real AdaBoost is used for feature selection and learning the stage classifiers from the training images. Experimental results show that our approach can detect people with both efficiency and accuracy.
Keywords :
edge detection; feature extraction; integral equations; learning (artificial intelligence); object detection; statistical analysis; traffic engineering computing; Real AdaBoost algorithm; boosted cascade structure; edge-orientation feature; feature selection; integral-histogram technique; pedestrian detection method; rectangle feature; Computer vision; Detectors; Face detection; Head; Histograms; Humans; Image edge detection; Motion detection; Object detection; Support vector machines;
Conference_Titel :
Intelligent Information Hiding and Multimedia Signal Processing, 2007. IIHMSP 2007. Third International Conference on
Conference_Location :
Kaohsiung
Print_ISBN :
978-0-7695-2994-1
DOI :
10.1109/IIHMSP.2007.4457735