Robust Low-Light Human Pose Estimation through Illumination-Texture Modulation

TL;DR

This paper introduces a frequency-based approach that enhances low-light images by separately processing low- and high-frequency components, leading to improved human pose estimation under challenging lighting conditions.

Contribution

It presents a novel frequency-based framework that selectively enhances image components, improving robustness and accuracy in low-light human pose estimation tasks.

Findings

Outperforms state-of-the-art methods in low-light scenarios

Effectively enhances semantic and texture information

Demonstrates robustness across various challenging conditions

Abstract

As critical visual details become obscured, the low visibility and high ISO noise in extremely low-light images pose a significant challenge to human pose estimation. Current methods fail to provide high-quality representations due to reliance on pixel-level enhancements that compromise semantics and the inability to effectively handle extreme low-light conditions for robust feature learning. In this work, we propose a frequency-based framework for low-light human pose estimation, rooted in the "divide-and-conquer" principle. Instead of uniformly enhancing the entire image, our method focuses on task-relevant information. By applying dynamic illumination correction to the low-frequency components and low-rank denoising to the high-frequency components, we effectively enhance both the semantic and texture information essential for accurate pose estimation. As a result, this targeted enhancement method results in robust, high-quality representations, significantly improving pose estimation performance. Extensive experiments demonstrating its superiority over state-of-the-art methods in various challenging low-light scenarios.