Pedestrian Detection Based on Multi-Stage Unsupervised Learning

In order to implement effective detection and utilize large numbers of unlabeled samples,a pedestrian detection method based on Unsupervised learning was presented.We apply deep learning to human detection to acquire pedestrian features with unlabeled data set.The detection method uses unsupervised convolution sparse auto-encoders to train features at all levels from the data set,then trains classifier with end-to-end supervised method.Additionally,we fine-tune the features in a supervised way.Experiments show that the method approach an state-of-art result on all data set.