LFNAT 2023 Challenge on Light Field Depth Estimation: Methods and Results

This paper reviews the 1st LFNAT challenge on light field depth estimation, which aims at predicting disparity information of central view image in a light field (i.e., pixel offset between central view image and adjacent view image). Compared to multi-view stereo matching, light field depth estimation emphasizes efficient utilization of the 2D angular information from multiple regularly varying views. This challenge specifies UrbanLF [20] light field dataset as the sole data source. There are two phases in total: submission phase and final evaluation phase, in which 75 registered participants successfully submit their predicted results in the first phase and 7 eligible teams compete in the second phase. The performance of all submissions is carefully reviewed and shown in this paper as a new standard for the current state-of-the-art in light field depth estimation. Moreover, the implementation details of these methods are also provided to stimulate related advanced research.

1. Introduction

Over the past few years, light field (LF) has gradually developed into one of the mainstream research areas of computer vision. Benefiting from the potential capability of additional directional information, LF simultaneously records both spatial information and angular information of all light rays. In addition, with the advent of plenoptic cameras, LF acquisition is greatly simplified and can be obtained in a single shot. Therefore, a large variety of research fields try to use LF instead of single image or video sequence for superior performance by mining internal structural information.