Point upsampling networks for single-photon sensing

TL;DR

This paper introduces a novel point upsampling network designed for single-photon sensing, significantly improving point cloud density and accuracy while reducing distortion, thus enhancing the practical utility of ultra-sensitive imaging.

Contribution

It is the first to establish an upsampling framework specifically for single-photon sensing, integrating advanced modules for improved reconstruction and robustness.

Findings

High reconstruction accuracy demonstrated on standard datasets

Robustness to distortion noise confirmed through experiments

Generated point clouds are visually consistent and detail-preserving

Abstract

Single-photon sensing has generated great interest as a prominent technique of long-distance and ultra-sensitive imaging, however, it tends to yield sparse and spatially biased point clouds, thus limiting its practical utility. In this work, we propose using point upsampling networks to increase point density and reduce spatial distortion in single-photon point cloud. Particularly, our network is built on the state space model which integrates a multi-path scanning mechanism to enrich spatial context, a bidirectional Mamba backbone to capture global geometry and local details, and an adaptive upsample shift module to correct offset-induced distortions. Extensive experiments are implemented on commonly-used datasets to confirm its high reconstruction accuracy and strong robustness to the distortion noise, and also on real-world data to demonstrate that our model is able to generate visually consistent, detail-preserving, and noise suppressed point clouds. Our work is the first to establish the upsampling framework for single-photon sensing, and hence opens a new avenue for single-photon sensing and its practical applications in the downstreaming tasks.