Self-supervised Depth Denoising Using Lower- and Higher-quality RGB-D sensors

TL;DR

This paper introduces a self-supervised deep learning method to improve the quality of low-grade depth sensors by leveraging synchronized high-quality RGB-D data, enhancing 3D reconstruction and tracking.

Contribution

It presents a novel self-supervised approach that aligns and uses high-quality depth data to denoise and refine low-quality depth measurements without requiring paired training data.

Findings

Outperforms existing filtering and deep denoising methods.

Results show improved 3D reconstruction detail and tracking accuracy.

Effective in real-world applications with consumer-grade sensors.

Abstract

Consumer-level depth cameras and depth sensors embedded in mobile devices enable numerous applications, such as AR games and face identification. However, the quality of the captured depth is sometimes insufficient for 3D reconstruction, tracking and other computer vision tasks. In this paper, we propose a self-supervised depth denoising approach to denoise and refine depth coming from a low quality sensor. We record simultaneous RGB-D sequences with unzynchronized lower- and higher-quality cameras and solve a challenging problem of aligning sequences both temporally and spatially. We then learn a deep neural network to denoise the lower-quality depth using the matched higher-quality data as a source of supervision signal. We experimentally validate our method against state-of-the-art filtering-based and deep denoising techniques and show its application for 3D object reconstruction tasks where our approach leads to more detailed fused surfaces and better tracking.