Architecture of asymmetrical pyramid non‐local block (APNB). H and W
![Architecture of asymmetrical pyramid non‐local block (APNB). H and W](https://www.researchgate.net/publication/351474964/figure/fig2/AS:11431281179148836@1691158726741/Architecture-of-asymmetrical-pyramid-non-local-block-APNB-H-and-W-are-the-height-and.png)
![](https://ars.els-cdn.com/content/image/1-s2.0-S0031320321003071-gr4.jpg)
ars.els-cdn.com/content/image/1-s2.0-S003132032100
![](https://ietresearch.onlinelibrary.wiley.com/cms/asset/00c859a2-1f20-47c5-be58-3ef6593208a3/itr212066-fig-0006-m.jpg)
Multi‐scale pedestrian detection based on self‐attention and adaptively spatial feature fusion - Wang - 2021 - IET Intelligent Transport Systems - Wiley Online Library
![](https://ietresearch.onlinelibrary.wiley.com/cms/asset/4483ed2b-eb7e-48d0-994d-b325ec417576/itr212066-fig-0004-m.jpg)
Multi‐scale pedestrian detection based on self‐attention and adaptively spatial feature fusion - Wang - 2021 - IET Intelligent Transport Systems - Wiley Online Library
![](https://d3i71xaburhd42.cloudfront.net/fd8b8a3d3dc059122971849a1e4f89b1d676bb04/8-Figure8-1.png)
PDF] Automatic Building Extraction From High-Resolution Aerial Imagery via Fully Convolutional Encoder-Decoder Network With Non-Local Block
![](https://www.researchgate.net/publication/363676033/figure/fig1/AS:11431281085220918@1663681216731/A-residual-dense-convolutional-block-is-composed-of-dense-concatenations-based-on_Q320.jpg)
A residual dense convolutional block is composed of dense
![](https://www.mdpi.com/sensors/sensors-23-03190/article_deploy/html/images/sensors-23-03190-g002-550.jpg)
Sensors, Free Full-Text
![](https://d3i71xaburhd42.cloudfront.net/10c2250aeab4d4cbdba07293da95a513dcdf85ee/4-Figure4-1.png)
PDF] Asymmetric Non-Local Neural Networks for Semantic Segmentation
![](https://www.mdpi.com/remotesensing/remotesensing-14-04600/article_deploy/html/images/remotesensing-14-04600-g001.png)
Remote Sensing, Free Full-Text
![](https://media.springernature.com/m685/springer-static/image/art%3A10.1007%2Fs00500-024-09638-7/MediaObjects/500_2024_9638_Fig5_HTML.png)
TCNet: tensor and covariance attention network for semantic segmentation