Pedestrian Detection in Thermal Images using Saliency Maps
This addresses the problem of pedestrian detection in thermal images for applications like surveillance or autonomous vehicles, offering an incremental improvement by eliminating the need for paired color images.
The paper tackles pedestrian detection in thermal images, which perform poorly during daytime, by augmenting thermal images with saliency maps as an attention mechanism, resulting in an absolute reduction of miss rate by 13.4% in day and 19.4% in night images over the baseline.
Thermal images are mainly used to detect the presence of people at night or in bad lighting conditions, but perform poorly at daytime. To solve this problem, most state-of-the-art techniques employ a fusion network that uses features from paired thermal and color images. Instead, we propose to augment thermal images with their saliency maps, to serve as an attention mechanism for the pedestrian detector especially during daytime. We investigate how such an approach results in improved performance for pedestrian detection using only thermal images, eliminating the need for paired color images. For our experiments, we train the Faster R-CNN for pedestrian detection and report the added effect of saliency maps generated using static and deep methods (PiCA-Net and R3-Net). Our best performing model results in an absolute reduction of miss rate by 13.4% and 19.4% over the baseline in day and night images respectively. We also annotate and release pixel level masks of pedestrians on a subset of the KAIST Multispectral Pedestrian Detection dataset, which is a first publicly available dataset for salient pedestrian detection.