.png){kind=icon}
.png){kind=icon}
Return
Rethinking U-Net architecture in medical imaging: Advancing the efficient and interpretable UKAN-CBAM framework for colorectal polyp segmentation
.png){kind=icon}
OA
AI
M
F
M
M
J
DOI：10.1016/j.artmed.2026.103352.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• UKAN-CBAM model for polyp segmentation on Kvasir-SEG dataset • Obtains 93.80% Dice, 89.18% IoU, 96.21% accuracy and 42 FPS • Delivers fast predictions (122.272 ms) with 5.214 GFLOPs and 14.29 M parameters • Demonstrates strong generalization on benchmark polyp image and video datasets • Uses feature maps, heatmaps, and Grad-CAM to interpret decision-making process
Keywords:
Colorectal cancer
Colorectal polyps
KANs (Kolmogorov-Arnold networks)
UKAN (U-Net with KAN)
CBAM (Convolutional block attention module)
UKAN-CBAM
Kvasir-SEG
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
6.2
Papers:
2.5K
Citations:
7.8K
Organization
Cited Papers