Two years have passed since the first workshop on Light Fields for Computer Vision at ECCV 2014, during which the field has advanced substantially. There have been breakthroughs in the differential theory of light fields allowing to better analyze non-Lambertian scenes, correct occlusion handling, multiple layer separation, saliency detection, tracking, material classification, and many more application areas. In particular, 3D reconstruction using light field data has made tremendous progress in accuracy and robustness, to the point that established benchmarks now lack in challenge and quality to drive further progress in the field. We believe it is therefore time to get together again and come up with joint ideas to push the limits of light field analysis, and aim to invite researchers from academia as well as industry to discuss promising future directions. Our workshop will focus on two main parts. First, it will be a venue to showcase the most recent research as well as current trends in technology. Second, we aim at achieving a quantitative valuation of state-of-the-art in depth reconstruction, and will discuss the requirements on novel benchmarks to measure progress and drive competition and advance in different areas.