Learnable Fractional Reaction-Diffusion Dynamics for Under-Display ToF Imaging and Beyond

TL;DR

This paper introduces LFRD2, a hybrid neural-physical model that enhances under-display ToF imaging by addressing signal degradation through fractional reaction-diffusion dynamics and continuous convolution, leading to improved depth sensing accuracy.

Contribution

The paper proposes a novel learnable fractional reaction-diffusion framework combining neural networks with physical modeling for depth enhancement in ToF imaging.

Findings

Effective depth refinement demonstrated on four benchmark datasets.

Outperforms existing methods in depth restoration quality.

Code is publicly available for reproducibility.

Abstract

Under-display ToF imaging aims to achieve accurate depth sensing through a ToF camera placed beneath a screen panel. However, transparent OLED (TOLED) layers introduce severe degradations-such as signal attenuation, multi-path interference (MPI), and temporal noise-that significantly compromise depth quality. To alleviate this drawback, we propose Learnable Fractional Reaction-Diffusion Dynamics (LFRD2), a hybrid framework that combines the expressive power of neural networks with the interpretability of physical modeling. Specifically, we implement a time-fractional reaction-diffusion module that enables iterative depth refinement with dynamically generated differential orders, capturing long-term dependencies. In addition, we introduce an efficient continuous convolution operator via coefficient prediction and repeated differentiation to further improve restoration quality. Experiments on four benchmark datasets demonstrate the effectiveness of our approach. The code is publicly available at https://github.com/wudiqx106/LFRD2.