Unsupervised Light Field Depth Estimation via Multi-view Feature Matching with Occlusion Prediction
This work addresses the costly need for labeled depth data in light field applications, offering an unsupervised solution with improved accuracy and robustness.
The authors tackled the problem of depth estimation from light field images without requiring labeled training data by proposing an unsupervised framework with multi-view feature matching and occlusion prediction. Their method achieved superior performance on both dense and sparse light field images and demonstrated better robustness and generalization on real-world data compared to other methods.
Depth estimation from light field (LF) images is a fundamental step for numerous applications. Recently, learning-based methods have achieved higher accuracy and efficiency than the traditional methods. However, it is costly to obtain sufficient depth labels for supervised training. In this paper, we propose an unsupervised framework to estimate depth from LF images. First, we design a disparity estimation network (DispNet) with a coarse-to-fine structure to predict disparity maps from different view combinations. It explicitly performs multi-view feature matching to learn the correspondences effectively. As occlusions may cause the violation of photo-consistency, we introduce an occlusion prediction network (OccNet) to predict the occlusion maps, which are used as the element-wise weights of photometric loss to solve the occlusion issue and assist the disparity learning. With the disparity maps estimated by multiple input combinations, we then propose a disparity fusion strategy based on the estimated errors with effective occlusion handling to obtain the final disparity map with higher accuracy. Experimental results demonstrate that our method achieves superior performance on both the dense and sparse LF images, and also shows better robustness and generalization on the real-world LF images compared to the other methods.