Semantic Segmentation using Light Attention Mechanism

Yuki Hiramatsu, Kazuhiro Hotta

Abstract

Semantic segmentation using convolutional neural networks (CNN) can be applied to various fields such as automatic driving. Semantic segmentation is pixel-wise class classification, and various methods using CNN have been proposed. We introduce a light attention mechanism to the encoder-decoder network. The network that introduced a light attention mechanism pays attention to features extracted during training, emphasizes the features judged to be effective for training and suppresses the features judged to be irrelevant for each pixel. As a result, training can be performed by focusing on only necessary features. We evaluated the proposed method using the CamVid dataset and obtained higher accuracy than conventional segmentation methods.

Download


Paper Citation


in Harvard Style

Hiramatsu Y. and Hotta K. (2020). Semantic Segmentation using Light Attention Mechanism.In Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 4: VISAPP, ISBN 978-989-758-402-2, pages 622-625. DOI: 10.5220/0009347206220625


in Bibtex Style

@conference{visapp20,
author={Yuki Hiramatsu and Kazuhiro Hotta},
title={Semantic Segmentation using Light Attention Mechanism},
booktitle={Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 4: VISAPP,},
year={2020},
pages={622-625},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0009347206220625},
isbn={978-989-758-402-2},
}


in EndNote Style

TY - CONF

JO - Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 4: VISAPP,
TI - Semantic Segmentation using Light Attention Mechanism
SN - 978-989-758-402-2
AU - Hiramatsu Y.
AU - Hotta K.
PY - 2020
SP - 622
EP - 625
DO - 10.5220/0009347206220625