End-to-end people detection in crowded scenes

Abstract

Current people detectors operate either by scanning an image in a sliding window fashion or by classifying a discrete set of proposals. We propose a model that is based on decoding an image into a set of people detections. Our system takes an image as input and directly outputs a set of distinct detection hypotheses. Because we generate predictions jointly, common post-processing steps such as non-maximum suppression are unnecessary. We use a recurrent LSTM layer for sequence generation and train our model end-to-end with a new loss function that operates on sets of detections. We demonstrate the effectiveness of our approach on the challenging task of detecting people in crowded scenes.

9 pages, 7 figures. Submitted to NIPS 2015. Supplementary material video: http://www.youtube.com/watch?v=QeWl0h3kQ24

References in corpus (4)

• Caffe: Convolutional Architecture for Fast Feature Embedding
• An Empirical Evaluation of Deep Learning on Highway Driving
• Deep Visual-Semantic Alignments for Generating Image Descriptions
• Filtered Channel Features for Pedestrian Detection

Cited by in corpus (7)

• Semantic Instance Segmentation with a Discriminative Loss Function
• Switching Convolutional Neural Network for Crowd Counting
• A Computer Vision System to Localize and Classify Wastes on the Streets
• Learning non-maximum suppression
• Recurrent Filter Learning for Visual Tracking
• Sequential Person Recognition in Photo Albums with a Recurrent Network
• A Video Analysis Method on Wanfang Dataset via Deep Neural Network