Causal Mode Multiplexer: A Novel Framework for Unbiased Multispectral Pedestrian Detection

TL;DR

This paper introduces a causal framework called Causal Mode Multiplexer to improve multispectral pedestrian detection by addressing modality bias, demonstrating better generalization across diverse datasets including a newly created ROTX-MP dataset.

Contribution

The paper proposes the Causal Mode Multiplexer framework that learns causal relationships to mitigate modality bias in multispectral pedestrian detection, and introduces a new dataset ROTX-MP for evaluation.

Findings

CMM improves generalization on existing datasets

CMM performs well on the new ROTX-MP dataset

The new dataset highlights modality bias issues

Abstract

RGBT multispectral pedestrian detection has emerged as a promising solution for safety-critical applications that require day/night operations. However, the modality bias problem remains unsolved as multispectral pedestrian detectors learn the statistical bias in datasets. Specifically, datasets in multispectral pedestrian detection mainly distribute between ROTO (day) and RXTO (night) data; the majority of the pedestrian labels statistically co-occur with their thermal features. As a result, multispectral pedestrian detectors show poor generalization ability on examples beyond this statistical correlation, such as ROTX data. To address this problem, we propose a novel Causal Mode Multiplexer (CMM) framework that effectively learns the causalities between multispectral inputs and predictions. Moreover, we construct a new dataset (ROTX-MP) to evaluate modality bias in multispectral pedestrian detection. ROTX-MP mainly includes ROTX examples not presented in previous datasets. Extensive experiments demonstrate that our proposed CMM framework generalizes well on existing datasets (KAIST, CVC-14, FLIR) and the new ROTX-MP. We will release our new dataset to the public for future research.