IACT: Intensive Attention in Convolution-Transformer Network for Facial Landmark Localization

Zhanyu Gao, Kai Chen, Dahai Yu

2023

Abstract

Recently, the facial landmarks localization tasks based on deep learning methods have achieved promising results, but they ignore the global context information and long-range relationship among the landmarks. To address this issue, we propose a parallel multi-branch architecture combining convolutional blocks and transformer layer for facial landmarks localization named Intensive Attention in the Convolutional Vision Transformer Network (IACT), which has the advantages of capturing detailed features and gathering global dynamic attention weights. To further improve the performance, the Intensive Attention mechanism is incorporated with the Convolution-Transformer Network, which includes Multi-head Spatial attention, Feature attention, the Channel attention. In addition, we present a novel loss function named Smooth Wing Loss that fills the gap in the gradient discontinuity of the Adaptive Wing loss, resulting in better convergence. Our IACT can achieve state-of-the-art performance on WFLW, 300W, and COFW datasets with 4.04, 2.82 and 3.12 in Normalized Mean Error.

Paper Citation

in Harvard Style

Gao Z., Chen K. and Yu D. (2023). IACT: Intensive Attention in Convolution-Transformer Network for Facial Landmark Localization. In Proceedings of the 18th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2023) - Volume 4: VISAPP; ISBN 978-989-758-634-7, SciTePress, pages 401-410. DOI: 10.5220/0011740900003417

in Bibtex Style

@conference{visapp23,
author={Zhanyu Gao and Kai Chen and Dahai Yu},
title={IACT: Intensive Attention in Convolution-Transformer Network for Facial Landmark Localization},
booktitle={Proceedings of the 18th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2023) - Volume 4: VISAPP},
year={2023},
pages={401-410},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011740900003417},
isbn={978-989-758-634-7},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 18th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2023) - Volume 4: VISAPP
TI - IACT: Intensive Attention in Convolution-Transformer Network for Facial Landmark Localization
SN - 978-989-758-634-7
AU - Gao Z.
AU - Chen K.
AU - Yu D.
PY - 2023
SP - 401
EP - 410
DO - 10.5220/0011740900003417
PB - SciTePress