Channel attention module.

<div><p>This paper presents a novel method for improving semantic segmentation performance in computer vision tasks. Our approach utilizes an enhanced UNet architecture that leverages an improved ResNet50 backbone. We replace the last layer of ResNet50 with deformable convolution to enha...

Full description

Saved in:

Bibliographic Details
Main Author:	Effat Sahragard (20570917) (author)
Other Authors:	Hassan Farsi (20570920) (author), Sajad Mohamadzadeh (20570923) (author)
Published:	2025
Subjects:	Molecular Biology Science Policy Space Science Biological Sciences not elsewhere classified Information Systems not elsewhere classified newly introduced wildpass investigate two types enhance feature representation computer vision tasks advancing semantic segmentation enhanced unet architecture enhanced unet algorithm efficient channel attention improved resnet50 backbone lka module outperforms incorporating attention attention mechanisms attention mechanism eca module xlink "> stanford dataset segmentation performance results indicate results demonstrate proposed method paper presents novel method lka modules level features last layer including stanford finding highlights encoding path deformable convolution decoding path decoder path concatenating low capture multi art results approach utilizes achieving state
Tags:	Add Tag No Tags, Be the first to tag this record!

Channel attention module.

Similar Items