A visual window convolutional attention mechanism based building extraction method for remote sensing images

Han Wang; Jingwen Li; Wenkang Tang

doi:10.1117/12.3005819

10 October 2023 A visual window convolutional attention mechanism based building extraction method for remote sensing images

Han Wang, Jingwen Li, Wenkang Tang

Author Affiliations +

Proceedings Volume 12799, Third International Conference on Advanced Algorithms and Signal Image Processing (AASIP 2023); 127995A (2023) https://doi.org/10.1117/12.3005819
Event: 3rd International Conference on Advanced Algorithms and Signal Image Processing (AASIP 2023), 2023, Kuala Lumpur, Malaysia

Abstract

To solve the problem of low accuracy and slow speed of building extraction due to complex semantic information of remote sensing image, Visual Window Convolutional Attention-Net(VWCA-Net) is constructed in this paper. VWCA-Net is based on an encoding-decoding framework, with the ResNet-50 used as the encoding backbone network in the encoding phase; the Visual Window Convolutional Attention Fusion (VWCAF) module is used to fuse shallow and deep features in the decoding phase. The VWCAF module first filters the feature maps containing buildings and generates Mask by windowing the sliced deep feature maps and fusing visual features, and then fuses the weights of CBAM to remove the features that do not contain the building feature maps, thus making the building boundary segmentation more accurate and faster. The experiment show that the Recall and mIoU of VWCA-Net on WHU dataset are better than SegNeXt network by 0.87% and 0.89%, respectively, but the computational effort is reduced by 17.38 FLOPs/G, which is a more effective and practical application method for building extraction.

(2023) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Han Wang, Jingwen Li, and Wenkang Tang "A visual window convolutional attention mechanism based building extraction method for remote sensing images", Proc. SPIE 12799, Third International Conference on Advanced Algorithms and Signal Image Processing (AASIP 2023), 127995A (10 October 2023); https://doi.org/10.1117/12.3005819

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available