Pedestrian Detection with Unsupervised Multi-Stage Feature Learning

TL;DR

This paper introduces a novel pedestrian detection approach that combines multi-stage feature learning, skip connections, and unsupervised convolutional sparse coding pre-training, achieving state-of-the-art results.

Contribution

It presents a new deep learning model with multi-stage features and unsupervised pre-training for improved pedestrian detection performance.

Findings

Achieved state-of-the-art results on major pedestrian datasets.

Demonstrated effectiveness of unsupervised pre-training with convolutional sparse coding.

Improved integration of global and local features through skip connections.

Abstract

Pedestrian detection is a problem of considerable practical interest. Adding to the list of successful applications of deep learning methods to vision, we report state-of-the-art and competitive results on all major pedestrian datasets with a convolutional network model. The model uses a few new twists, such as multi-stage features, connections that skip layers to integrate global shape information with local distinctive motif information, and an unsupervised method based on convolutional sparse coding to pre-train the filters at each stage.