New approach for human detection in images using histograms of oriented gradients

Histograms of Oriented Gradients (HoG) is one of the most used descriptors in human detection. Although it has good performance compared to other descriptors in the area, if the size and number of images increase, the dimension of the descriptor vectors would become extremely large and therefore makes the training process computationally complex. To overcome this, in this paper a human detection method based on bag-of-features model is represented. Visual words are patches of pictures described with HoG and then clustered using K-means algorithm. To highlight the most important visual words, a weighting method could be applied to the descriptor vectors. Here we used Term Frequency-Inverse Document Frequency (Tf_Idf) which has been used in document classification. In the proposed approach, Support Vector Machine (SVM) is used as the binary classifier. We applied our proposed method to the MIT and INRIA datasets and compared the performance of our algorithm with a similar method in the literature. The results of our experiments show that our method performs at least as well as other available methods.