A learning-based method using Epipolar Geometry for light field depth estimation

摘要

A novel method is proposed in this paper for light field depth estimation by using a convolutional neural network. Manyapproaches have been proposed to make light field depth estimation, while most of them have a contradiction betweenaccuracy and runtime. In order to solve this problem, we proposed a method which can get more accurate light field depthestimation results with faster speed. First, the light field data is augmented by proposed method considering the light fieldgeometry. Because of the large amount of the light field data, the number of images needs to be reduced appropriately toimprove the operation speed, while maintaining the confidence of the estimation. Next, light field images are inputted intoour network after data augmentation. The features of the images are extracted during the process, which could be used tocalculate the disparity value. Finally, our network can generate an accurate depth map from the input light field image aftertraining. Using this accurate depth map, the 3D structure in real world could be accurately reconstructed. Our method isverified by the HCI 4D Light Field Benchmark and real-world light field images captured with a Lytro light field camera.