I. Introduction
Light Field (LF) [1], [2] depth estimation is a fundamental task in computer vision with many subsequent applications, such as refocusing [1], [3], super-resolution [4], [5], [6], [7], [8], [9], view synthesis [10], [11], [12], 3D reconstruction [13], and virtual reality [14]. The 4D LF records abundant spatial and angular information in the scene, from which the scene depth can be estimated. Based on the characteristics of light field images, traditional methods [15], [16], [17], [18] and learning-based methods [19], [20], [21], [22], [23], [24] have been proposed for light field image depth estimation.