To improve segmentation performance, a novel neural network architecture (termed DFCN-DCRF) is proposed, which combines an RGB-D fully convolutional neural network (DFCN) with a depth-sensitive fully-connected conditional random field (DCRF). First, a DFCN architecture which fuses depth information into the early layers and applies dilated convolution for later contextual reasoning is designed. Then, a depth-sensitive fully-connected conditional random field (DCRF) is proposed and combined with the previous DFCN to refine the preliminary result. Comparative experiments show that the proposed DFCN-DCRF has the best performance compared with most state-of-the-art methods.
@article{arxiv.1705.07383,
title = {Incorporating Depth into both CNN and CRF for Indoor Semantic Segmentation},
author = {Jindong Jiang and Zhijun Zhang and Yongqian Huang and Lunan Zheng},
journal= {arXiv preprint arXiv:1705.07383},
year = {2018}
}