GRAM: Enhancing Medical Image Segmentation with Global-local and Region Attention Module

Chia Yu Chu; Cheng Hung Lin

doi:10.1109/ICCE-Taiwan66881.2025.11207985

GRAM: Enhancing Medical Image Segmentation with Global-local and Region Attention Module

Chia Yu Chu^*
, Cheng Hung Lin

^*Corresponding author for this work

Department of Electrical Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Accurate and efficient segmentation of medical images remains a critical challenge in computer-aided diagnosis, particularly in dermatology. We propose GRAM (Global-local and Region Attention Module), a novel hybrid attention framework that combines global contextual awareness with fine-grained spatial detail refinement. GRAM effectively mitigates challenges such as boundary ambiguity and loss of contextual information, leading to improved segmentation accuracy. Evaluations on the ISIC 2016 dataset demonstrate the effectiveness of GRAM, achieving notable improvements in IoU and Dice scores.

Original language	English
Title of host publication	ICCE-Taiwan 2025 - 12th IEEE International Conference on Consumer Electronics - Taiwan
Subtitle of host publication	Generative AI in Innovative Consumer Technology, Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	57-58
Number of pages	2
ISBN (Electronic)	9798331587413
DOIs	https://doi.org/10.1109/ICCE-Taiwan66881.2025.11207985
Publication status	Published - 2025
Event	12th IEEE International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2025 - Kaohsiung, Taiwan Duration: 2025 Jul 16 → 2025 Jul 18

Publication series

Name	ICCE-Taiwan 2025 - 12th IEEE International Conference on Consumer Electronics - Taiwan: Generative AI in Innovative Consumer Technology, Proceedings

Conference

Conference	12th IEEE International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2025
Country/Territory	Taiwan
City	Kaohsiung
Period	2025/07/16 → 2025/07/18

Keywords

Attention Mechanisms
Global-Local and Region Attention Module
lesion segmentation
transformer

ASJC Scopus subject areas

Human-Computer Interaction
Electrical and Electronic Engineering
Media Technology
Modelling and Simulation
Instrumentation

Access to Document

10.1109/ICCE-Taiwan66881.2025.11207985

Cite this

Chu, C. Y., & Lin, C. H. (2025). GRAM: Enhancing Medical Image Segmentation with Global-local and Region Attention Module. In ICCE-Taiwan 2025 - 12th IEEE International Conference on Consumer Electronics - Taiwan: Generative AI in Innovative Consumer Technology, Proceedings (pp. 57-58). (ICCE-Taiwan 2025 - 12th IEEE International Conference on Consumer Electronics - Taiwan: Generative AI in Innovative Consumer Technology, Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICCE-Taiwan66881.2025.11207985

GRAM: Enhancing Medical Image Segmentation with Global-local and Region Attention Module. / Chu, Chia Yu; Lin, Cheng Hung.
ICCE-Taiwan 2025 - 12th IEEE International Conference on Consumer Electronics - Taiwan: Generative AI in Innovative Consumer Technology, Proceedings. Institute of Electrical and Electronics Engineers Inc., 2025. p. 57-58 (ICCE-Taiwan 2025 - 12th IEEE International Conference on Consumer Electronics - Taiwan: Generative AI in Innovative Consumer Technology, Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Chu, CY & Lin, CH 2025, GRAM: Enhancing Medical Image Segmentation with Global-local and Region Attention Module. in ICCE-Taiwan 2025 - 12th IEEE International Conference on Consumer Electronics - Taiwan: Generative AI in Innovative Consumer Technology, Proceedings. ICCE-Taiwan 2025 - 12th IEEE International Conference on Consumer Electronics - Taiwan: Generative AI in Innovative Consumer Technology, Proceedings, Institute of Electrical and Electronics Engineers Inc., pp. 57-58, 12th IEEE International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2025, Kaohsiung, Taiwan, 2025/07/16. https://doi.org/10.1109/ICCE-Taiwan66881.2025.11207985

Chu CY, Lin CH. GRAM: Enhancing Medical Image Segmentation with Global-local and Region Attention Module. In ICCE-Taiwan 2025 - 12th IEEE International Conference on Consumer Electronics - Taiwan: Generative AI in Innovative Consumer Technology, Proceedings. Institute of Electrical and Electronics Engineers Inc. 2025. p. 57-58. (ICCE-Taiwan 2025 - 12th IEEE International Conference on Consumer Electronics - Taiwan: Generative AI in Innovative Consumer Technology, Proceedings). doi: 10.1109/ICCE-Taiwan66881.2025.11207985

Chu, Chia Yu ; Lin, Cheng Hung. / GRAM : Enhancing Medical Image Segmentation with Global-local and Region Attention Module. ICCE-Taiwan 2025 - 12th IEEE International Conference on Consumer Electronics - Taiwan: Generative AI in Innovative Consumer Technology, Proceedings. Institute of Electrical and Electronics Engineers Inc., 2025. pp. 57-58 (ICCE-Taiwan 2025 - 12th IEEE International Conference on Consumer Electronics - Taiwan: Generative AI in Innovative Consumer Technology, Proceedings).

@inproceedings{27d3f7e505e3477a97f22798cfb2f171,

title = "GRAM: Enhancing Medical Image Segmentation with Global-local and Region Attention Module",

abstract = "Accurate and efficient segmentation of medical images remains a critical challenge in computer-aided diagnosis, particularly in dermatology. We propose GRAM (Global-local and Region Attention Module), a novel hybrid attention framework that combines global contextual awareness with fine-grained spatial detail refinement. GRAM effectively mitigates challenges such as boundary ambiguity and loss of contextual information, leading to improved segmentation accuracy. Evaluations on the ISIC 2016 dataset demonstrate the effectiveness of GRAM, achieving notable improvements in IoU and Dice scores.",

keywords = "Attention Mechanisms, Global-Local and Region Attention Module, lesion segmentation, transformer",

author = "Chu, \{Chia Yu\} and Lin, \{Cheng Hung\}",

note = "Publisher Copyright: {\textcopyright} 2025 IEEE.; 12th IEEE International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2025 ; Conference date: 16-07-2025 Through 18-07-2025",

year = "2025",

doi = "10.1109/ICCE-Taiwan66881.2025.11207985",

language = "English",

series = "ICCE-Taiwan 2025 - 12th IEEE International Conference on Consumer Electronics - Taiwan: Generative AI in Innovative Consumer Technology, Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "57--58",

booktitle = "ICCE-Taiwan 2025 - 12th IEEE International Conference on Consumer Electronics - Taiwan",

}

TY - GEN

T1 - GRAM

T2 - 12th IEEE International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2025

AU - Chu, Chia Yu

AU - Lin, Cheng Hung

PY - 2025

Y1 - 2025

N2 - Accurate and efficient segmentation of medical images remains a critical challenge in computer-aided diagnosis, particularly in dermatology. We propose GRAM (Global-local and Region Attention Module), a novel hybrid attention framework that combines global contextual awareness with fine-grained spatial detail refinement. GRAM effectively mitigates challenges such as boundary ambiguity and loss of contextual information, leading to improved segmentation accuracy. Evaluations on the ISIC 2016 dataset demonstrate the effectiveness of GRAM, achieving notable improvements in IoU and Dice scores.

AB - Accurate and efficient segmentation of medical images remains a critical challenge in computer-aided diagnosis, particularly in dermatology. We propose GRAM (Global-local and Region Attention Module), a novel hybrid attention framework that combines global contextual awareness with fine-grained spatial detail refinement. GRAM effectively mitigates challenges such as boundary ambiguity and loss of contextual information, leading to improved segmentation accuracy. Evaluations on the ISIC 2016 dataset demonstrate the effectiveness of GRAM, achieving notable improvements in IoU and Dice scores.

KW - Attention Mechanisms

KW - Global-Local and Region Attention Module

KW - lesion segmentation

KW - transformer

UR - https://www.scopus.com/pages/publications/105022406260

UR - https://www.scopus.com/pages/publications/105022406260#tab=citedBy

U2 - 10.1109/ICCE-Taiwan66881.2025.11207985

DO - 10.1109/ICCE-Taiwan66881.2025.11207985

M3 - Conference contribution

AN - SCOPUS:105022406260

T3 - ICCE-Taiwan 2025 - 12th IEEE International Conference on Consumer Electronics - Taiwan: Generative AI in Innovative Consumer Technology, Proceedings

SP - 57

EP - 58

BT - ICCE-Taiwan 2025 - 12th IEEE International Conference on Consumer Electronics - Taiwan

PB - Institute of Electrical and Electronics Engineers Inc.

Y2 - 16 July 2025 through 18 July 2025

ER -

GRAM: Enhancing Medical Image Segmentation with Global-local and Region Attention Module

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this