Research on Pedestrian Detection Based on Faster R-CNN and Hippocampal Neural Network

Abstract

This paper use Faster-RCNN and hippocampal neural network algorithms to research. Firstly use convolutional neural network to extract the features of the input image, and then use Region Proposal Networks to extract the standard frame. Here we can judge whether there are objects in the standard frame and know the location of the standard box, then use the Non-Maximum Suppression to select the standard box, finally perform the classification operation and regression operation. The final classification network is the hippocampal neural network. The hippocampal neural network is a spatial structure model that mimics the hippocampus of human brain.