Pedestrian recognition method based on depth hierarchical feature representation

For feature representation of pedestrian recognition, a hybrid hierarchical feature representation method which combines representation ability of bag of words model and depth layered with learning adaptability is presented. This method first uses HOG local descriptor for local features extraction, and then encoding the feature by a depth of layered coding method, the layered coding method by spatial aggregating restricted Boltzmann machine (RBM). For each coding layer, we steer the unsupervised RBM learning and apply supervised fine-tuning to enhance the visual features representation in classification task. Finally, we learn high-level image feature representation by the positive and negative max pooling, and then classify with the linear support vector machine, feature extraction of depth architectures effectively improve the accuracy of subsequent recognition. Experimental results show that the proposed method has a high recognition rate.