Light Field Segmentation Using a Ray-Based Graph Structure

International audienceIn this paper, we introduce a novel graph representation forinteractive light eld segmentation using Markov Random Field (MRF).The greatest barrier to the adoption of MRF for light eld processing isthe large volume of input data. The proposed graph structure exploits theredundancy in the ray space in order to reduce the graph size, decreasingthe running time of MRF-based optimisation tasks. Concepts of free raysand ray bundles with corresponding neighbourhood relationships are de-ned to construct the simplied graph-based light eld representation.We then propose a light eld interactive segmentation algorithm usinggraph-cuts based on such ray space graph structure, that guarantees thesegmentation consistency across all views. Our experiments with severaldatasets show results that are very close to the ground truth, competingwith state of the art light eld segmentation methods in terms of accuracyand with a signicantly lower complexity. They also show that ourmethod performs well on both densely and sparsely sampled light fields