Abstract
We propose a method for depth estimation from light field data, based on afully convolutional neural network architecture. Our goal is to design apipeline which achieves highly accurate results for small- and wide-baselinelight fields. Since light field training data is scarce, all learning-basedapproaches use a small receptive field and operate on small disparity ranges.In order to work with wide-baseline light fields, we introduce the idea ofEPI-Shift: To virtually shift the light field stack which enables to retain asmall receptive field, independent of the disparity range. In this way, ourapproach "learns to think outside the box of the receptive field". Our networkperforms joint classification of integer disparities and regression ofdisparity-offsets. A U-Net component provides excellent long-range smoothing.EPI-Shift considerably outperforms the state-of-the-art learning-basedapproaches and is on par with hand-crafted methods. We demonstrate this on apublicly available, synthetic, small-baseline benchmark and on large-baselinereal-world recordings.