Learning Multi-Level Region Consistency with Dense Multi-Label Networks for Semantic Segmentation
Abstract
Semantic image segmentation is a fundamental task in image understanding. Per-pixel semantic labelling of an image benefits greatly from the ability to consider region consistency both locally and globally. However, many Fully Convolutional Network based methods do not impose such consistency, which may give rise to noisy and implausible predictions. We address this issue by proposing a dense multi-label network module that is able to encourage the region consistency at different levels. This simple but effective module can be easily integrated into any semantic segmentation systems. With comprehensive experiments, we show that the dense multi-label can successfully remove the implausible labels and clear the confusion so as to boost the performance of semantic segmentation systems.
Cite
Text
Shen et al. "Learning Multi-Level Region Consistency with Dense Multi-Label Networks for Semantic Segmentation." International Joint Conference on Artificial Intelligence, 2017. doi:10.24963/IJCAI.2017/377Markdown
[Shen et al. "Learning Multi-Level Region Consistency with Dense Multi-Label Networks for Semantic Segmentation." International Joint Conference on Artificial Intelligence, 2017.](https://mlanthology.org/ijcai/2017/shen2017ijcai-learning/) doi:10.24963/IJCAI.2017/377BibTeX
@inproceedings{shen2017ijcai-learning,
title = {{Learning Multi-Level Region Consistency with Dense Multi-Label Networks for Semantic Segmentation}},
author = {Shen, Tong and Lin, Guosheng and Shen, Chunhua and Reid, Ian D.},
booktitle = {International Joint Conference on Artificial Intelligence},
year = {2017},
pages = {2708-2714},
doi = {10.24963/IJCAI.2017/377},
url = {https://mlanthology.org/ijcai/2017/shen2017ijcai-learning/}
}