J-CaPA : Joint Channel and Pyramid Attention Improves Medical Image Segmentation

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Nizam, Marzia Binta, Zlateva, Marian, Davis, James
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866915033509789696
author Nizam, Marzia Binta
Zlateva, Marian
Davis, James
author_facet Nizam, Marzia Binta
Zlateva, Marian
Davis, James
contents Medical image segmentation is crucial for diagnosis and treatment planning. Traditional CNN-based models, like U-Net, have shown promising results but struggle to capture long-range dependencies and global context. To address these limitations, we propose a transformer-based architecture that jointly applies Channel Attention and Pyramid Attention mechanisms to improve multi-scale feature extraction and enhance segmentation performance for medical images. Increasing model complexity requires more training data, and we further improve model generalization with CutMix data augmentation. Our approach is evaluated on the Synapse multi-organ segmentation dataset, achieving a 6.9% improvement in Mean Dice score and a 39.9% improvement in Hausdorff Distance (HD95) over an implementation without our enhancements. Our proposed model demonstrates improved segmentation accuracy for complex anatomical structures, outperforming existing state-of-the-art methods.
format Preprint
id arxiv_https___arxiv_org_abs_2411_16568
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle J-CaPA : Joint Channel and Pyramid Attention Improves Medical Image Segmentation
Nizam, Marzia Binta
Zlateva, Marian
Davis, James
Computer Vision and Pattern Recognition
Medical image segmentation is crucial for diagnosis and treatment planning. Traditional CNN-based models, like U-Net, have shown promising results but struggle to capture long-range dependencies and global context. To address these limitations, we propose a transformer-based architecture that jointly applies Channel Attention and Pyramid Attention mechanisms to improve multi-scale feature extraction and enhance segmentation performance for medical images. Increasing model complexity requires more training data, and we further improve model generalization with CutMix data augmentation. Our approach is evaluated on the Synapse multi-organ segmentation dataset, achieving a 6.9% improvement in Mean Dice score and a 39.9% improvement in Hausdorff Distance (HD95) over an implementation without our enhancements. Our proposed model demonstrates improved segmentation accuracy for complex anatomical structures, outperforming existing state-of-the-art methods.
title J-CaPA : Joint Channel and Pyramid Attention Improves Medical Image Segmentation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2411.16568