A Novel Combination Feature HOG-LSS for Pedestrian Detection

Since the ability of various kinds of feature descriptor is different in pedestrian detection and selecting them is not always fathomed, the six common features are analyzed in theory and compared in experiments. It is expected to find a new feature with the strongest description ability from their pair-wise combinations. In experiments, INRIA database and Daimler database are selected as the sample set. Adaboost is regarded as classifier and the detection performance is evaluated by detection rate, false alarm rate and detection time. The results of these three indicators further prove that description ability of HOG-LSS feature is better than others.