M3FPolypSegNet ++: Adaptive Gaussian frequency decomposition with multi-task learning for polyp segmentation

doi:10.1016/j.neucom.2025.131134

M3FPolypSegNet ++: Adaptive Gaussian frequency decomposition with multi-task learning for polyp segmentation

2025-07-28

0

PRE

AI

DOI：10.1016/j.neucom.2025.131134

Original

Original request for help

Save

Abstract

En 中文

• We propose GAF-ASPP, adaptively separating features via learnable Gaussian filtering. • We adopt a Transformer-based encoder to capture global context. • M3FPolypSegNet ++ surpasses prior CNN, Transformer, and hybrid segmentation models.

Keywords:

GAF-ASPP

Transformer-based encoder

feature separation

global context

polyp segmentation

M3FPolypSegNet ++: Adaptive Gaussian frequency decomposition with multi-task learning for polyp segmentation

Abstract

Journal

Organization