Research Of Pedestrian Recognition Methods Based On Depth Feature Representation

Pedestrian detection and recognition technology has been an important application in the field of intelligent transportation and intelligent monitoring system. The technology in the past decades has made considerable development, but due to the influence of pedestrian appearance, complex background and other factors, pedestrian recognition technology is still open problems in computer vision field. In view of this problem, we proposes a novel method of pedestrian recognition, which is compared with the similar method on 3 pedestrian data sets, the result proves that the proposed method has a higher recognition rate.For feature representation of pedestrian recognition, a hybrid hierarchical feature representation method which combines representation ability of the bag of words model and depth layered with learning adaptability is presented. This method first uses HOG local descriptor gradient-based 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, the sparse and selective regularization are used for the unsupervised RBM learning and supervision fine-tuning is used to enhance the visual features representation in classification task. Finally, high-level image feature representation is obtained by the max pooling and space of Pyramid method, and then the linear support vector machine is used for pedestrian recognition, feature extraction of depth architecture improves effectively the accuracy of subsequent recognition. Experimental results show that the proposed method has a high recognition rate.In view of the limitation of the max pooling method, the dissertation uses the method of positive and negative max pooling for the pedestrian recognition. The positive and negative max pooling method overcomes the problem of non-negative coefficient of encoding in max pooling method, which further improves the recognition accuracy rate, but the real-time performance is still not effectively resolved.Finally, we summarizes the research work of dissertation, points out the problems and difficulties of pedestrian recognition, and looks forward to solving the relevant problems in the future.