Unsupervised Learning Based Focal Stack Camera Depth Estimation

We propose an unsupervised deep learning based method to estimate depth from focal stack camera images. On the NYU-v2 dataset, our method achieves much better depth estimation accuracy compared to single-image based methods.