Deep End-to-End Alignment and Refinement for Time-of-Flight RGB-D Module

TL;DR

This paper introduces a deep learning framework that jointly aligns and refines ToF depth data with RGB images, improving accuracy and quality for mobile depth sensing applications.

Contribution

It presents a novel end-to-end deep learning method for simultaneous alignment and refinement of ToF and RGB data, including a new dataset synthesis approach.

Findings

Achieves state-of-the-art ToF depth refinement results

Effective joint alignment and refinement via deep learning

Enhanced depth quality for mobile RGB-D systems

Abstract

Recently, it is increasingly popular to equip mobile RGB cameras with Time-of-Flight (ToF) sensors for active depth sensing. However, for off-the-shelf ToF sensors, one must tackle two problems in order to obtain high-quality depth with respect to the RGB camera, namely 1) online calibration and alignment; and 2) complicated error correction for ToF depth sensing. In this work, we propose a framework for jointly alignment and refinement via deep learning. First, a cross-modal optical flow between the RGB image and the ToF amplitude image is estimated for alignment. The aligned depth is then refined via an improved kernel predicting network that performs kernel normalization and applies the bias prior to the dynamic convolution. To enrich our data for end-to-end training, we have also synthesized a dataset using tools from computer graphics. Experimental results demonstrate the effectiveness of our approach, achieving state-of-the-art for ToF refinement.