Robust photon-efficient imaging using a pixel-wise residual shrinkage network

TL;DR

This paper introduces a novel pixel-wise residual shrinkage network that enhances photon-efficient 3D imaging by adaptively denoising high-noise LiDAR data, achieving superior accuracy and robustness in challenging conditions.

Contribution

It proposes a new deep learning model that adaptively thresholds and denoises LiDAR data, redefining the optimization as pixel-wise classification for improved depth estimation.

Findings

Outperforms existing methods on simulated and real datasets.

Maintains robust performance even at extreme noise levels (1:100 SNR).

Achieves more confident and accurate depth predictions.

Abstract

Single-photon light detection and ranging (LiDAR) has been widely applied to 3D imaging in challenging scenarios. However, limited signal photon counts and high noises in the collected data have posed great challenges for predicting the depth image precisely. In this paper, we propose a pixel-wise residual shrinkage network for photon-efficient imaging from high-noise data, which adaptively generates the optimal thresholds for each pixel and denoises the intermediate features by soft thresholding. Besides, redefining the optimization target as pixel-wise classification provides a sharp advantage in producing confident and accurate depth estimation when compared with existing research. Comprehensive experiments conducted on both simulated and real-world datasets demonstrate that the proposed model outperforms the state-of-the-arts and maintains robust imaging performance under different signal-to-noise ratios including the extreme case of 1:100.