LiteNet: A Lightweight Feature Fusion Model for Brain Tumor Classification

Abdul Haseeb Nizamani; Zhigang Chen; Ahsan Ahmed Nizamani; Ali M.A. Ibrahim

doi:10.1007/978-981-96-6960-8_10

LiteNet: A Lightweight Feature Fusion Model for Brain Tumor Classification

Abdul Haseeb Nizamani^*
, Zhigang Chen
, Ahsan Ahmed Nizamani
, Ali M.A. Ibrahim

^*Corresponding author for this work

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

Abstract

Medical image analysis is crucial for the accurate diagnosis and treatment of brain tumors in modern healthcare. This study introduces LiteNet, an efficient feature fusion model for brain tumor classification using MRI scans. Combining EfficientNetB0 and MobileNetV2, LiteNet integrates a Lightweight Feature Extraction Module (LEM) and a Feature Fusion Module (FFM) to enhance feature representation while maintaining computational efficiency. Evaluated on Figshare and Br35H datasets, LiteNet achieved accuracies of 99.02% and 99.83%, respectively. Precision, recall, and F1-Score of 99.89%, 98.88%, and 98.88% on Figshare, and consistent 99.83% on Br35H, indicate its robustness on diverse datasets. We utilized Grad-CAM visualization to provide interpretability by highlighting important regions contributing to classification decisions. These impressive results underscore LiteNet’s capability to accurately detect and classify brain tumors, making it a valuable tool for enhancing diagnostic accuracy in clinical settings, particularly in resource-limited environments.

Original language	English
Title of host publication	Neural Information Processing - 31st International Conference, ICONIP 2024, Proceedings
Editors	Mufti Mahmud, Maryam Doborjeh, Zohreh Doborjeh, Kevin Wong, Andrew Chi Sing Leung, M. Tanveer
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	132-147
Number of pages	16
ISBN (Print)	9789819669592
DOIs	https://doi.org/10.1007/978-981-96-6960-8_10
State	Published - 2025
Externally published	Yes
Event	31st International Conference on Neural Information Processing, ICONIP 2024 - Auckland, New Zealand Duration: 2 Dec 2024 → 6 Dec 2024

Publication series

Name	Communications in Computer and Information Science
Volume	2286 CCIS
ISSN (Print)	1865-0929
ISSN (Electronic)	1865-0937

Conference

Conference	31st International Conference on Neural Information Processing, ICONIP 2024
Country/Territory	New Zealand
City	Auckland
Period	2/12/24 → 6/12/24

Bibliographical note

Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

Keywords

Classification
EfficientNetB0
Feature fusion
Lightweight feature extraction
MobileNetV2
Transfer learning

ASJC Scopus subject areas

General Computer Science
General Mathematics

Access to Document

10.1007/978-981-96-6960-8_10

Cite this

Nizamani, A. H., Chen, Z., Nizamani, AA., & Ibrahim, A. M. A. (2025). LiteNet: A Lightweight Feature Fusion Model for Brain Tumor Classification. In M. Mahmud, M. Doborjeh, Z. Doborjeh, K. Wong, A. C. S. Leung, & M. Tanveer (Eds.), Neural Information Processing - 31st International Conference, ICONIP 2024, Proceedings (pp. 132-147). (Communications in Computer and Information Science; Vol. 2286 CCIS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-981-96-6960-8_10

Nizamani, Abdul Haseeb ; Chen, Zhigang ; Nizamani, Ahsan Ahmed et al. / LiteNet : A Lightweight Feature Fusion Model for Brain Tumor Classification. Neural Information Processing - 31st International Conference, ICONIP 2024, Proceedings. editor / Mufti Mahmud ; Maryam Doborjeh ; Zohreh Doborjeh ; Kevin Wong ; Andrew Chi Sing Leung ; M. Tanveer. Springer Science and Business Media Deutschland GmbH, 2025. pp. 132-147 (Communications in Computer and Information Science).

@inproceedings{58995f3a498f43d9af453ae2e2388e31,

title = "LiteNet: A Lightweight Feature Fusion Model for Brain Tumor Classification",

abstract = "Medical image analysis is crucial for the accurate diagnosis and treatment of brain tumors in modern healthcare. This study introduces LiteNet, an efficient feature fusion model for brain tumor classification using MRI scans. Combining EfficientNetB0 and MobileNetV2, LiteNet integrates a Lightweight Feature Extraction Module (LEM) and a Feature Fusion Module (FFM) to enhance feature representation while maintaining computational efficiency. Evaluated on Figshare and Br35H datasets, LiteNet achieved accuracies of 99.02\% and 99.83\%, respectively. Precision, recall, and F1-Score of 99.89\%, 98.88\%, and 98.88\% on Figshare, and consistent 99.83\% on Br35H, indicate its robustness on diverse datasets. We utilized Grad-CAM visualization to provide interpretability by highlighting important regions contributing to classification decisions. These impressive results underscore LiteNet{\textquoteright}s capability to accurately detect and classify brain tumors, making it a valuable tool for enhancing diagnostic accuracy in clinical settings, particularly in resource-limited environments.",

keywords = "Classification, EfficientNetB0, Feature fusion, Lightweight feature extraction, MobileNetV2, Transfer learning",

author = "Nizamani, \{Abdul Haseeb\} and Zhigang Chen and Ahsan Ahmed Nizamani and Ibrahim, \{Ali M.A.\}",

note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.; 31st International Conference on Neural Information Processing, ICONIP 2024 ; Conference date: 02-12-2024 Through 06-12-2024",

year = "2025",

doi = "10.1007/978-981-96-6960-8\_10",

language = "English",

isbn = "9789819669592",

series = "Communications in Computer and Information Science",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "132--147",

editor = "Mufti Mahmud and Maryam Doborjeh and Zohreh Doborjeh and Kevin Wong and Leung, \{Andrew Chi Sing\} and M. Tanveer",

booktitle = "Neural Information Processing - 31st International Conference, ICONIP 2024, Proceedings",

address = "Germany",

}

Nizamani, AH, Chen, Z, Nizamani, AA & Ibrahim, AMA 2025, LiteNet: A Lightweight Feature Fusion Model for Brain Tumor Classification. in M Mahmud, M Doborjeh, Z Doborjeh, K Wong, ACS Leung & M Tanveer (eds), Neural Information Processing - 31st International Conference, ICONIP 2024, Proceedings. Communications in Computer and Information Science, vol. 2286 CCIS, Springer Science and Business Media Deutschland GmbH, pp. 132-147, 31st International Conference on Neural Information Processing, ICONIP 2024, Auckland, New Zealand, 2/12/24. https://doi.org/10.1007/978-981-96-6960-8_10

LiteNet: A Lightweight Feature Fusion Model for Brain Tumor Classification. / Nizamani, Abdul Haseeb; Chen, Zhigang; Nizamani, Ahsan Ahmed et al.
Neural Information Processing - 31st International Conference, ICONIP 2024, Proceedings. ed. / Mufti Mahmud; Maryam Doborjeh; Zohreh Doborjeh; Kevin Wong; Andrew Chi Sing Leung; M. Tanveer. Springer Science and Business Media Deutschland GmbH, 2025. p. 132-147 (Communications in Computer and Information Science; Vol. 2286 CCIS).

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

TY - GEN

T1 - LiteNet

T2 - 31st International Conference on Neural Information Processing, ICONIP 2024

AU - Nizamani, Abdul Haseeb

AU - Chen, Zhigang

AU - Nizamani, Ahsan Ahmed

AU - Ibrahim, Ali M.A.

N1 - Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

PY - 2025

Y1 - 2025

N2 - Medical image analysis is crucial for the accurate diagnosis and treatment of brain tumors in modern healthcare. This study introduces LiteNet, an efficient feature fusion model for brain tumor classification using MRI scans. Combining EfficientNetB0 and MobileNetV2, LiteNet integrates a Lightweight Feature Extraction Module (LEM) and a Feature Fusion Module (FFM) to enhance feature representation while maintaining computational efficiency. Evaluated on Figshare and Br35H datasets, LiteNet achieved accuracies of 99.02% and 99.83%, respectively. Precision, recall, and F1-Score of 99.89%, 98.88%, and 98.88% on Figshare, and consistent 99.83% on Br35H, indicate its robustness on diverse datasets. We utilized Grad-CAM visualization to provide interpretability by highlighting important regions contributing to classification decisions. These impressive results underscore LiteNet’s capability to accurately detect and classify brain tumors, making it a valuable tool for enhancing diagnostic accuracy in clinical settings, particularly in resource-limited environments.

AB - Medical image analysis is crucial for the accurate diagnosis and treatment of brain tumors in modern healthcare. This study introduces LiteNet, an efficient feature fusion model for brain tumor classification using MRI scans. Combining EfficientNetB0 and MobileNetV2, LiteNet integrates a Lightweight Feature Extraction Module (LEM) and a Feature Fusion Module (FFM) to enhance feature representation while maintaining computational efficiency. Evaluated on Figshare and Br35H datasets, LiteNet achieved accuracies of 99.02% and 99.83%, respectively. Precision, recall, and F1-Score of 99.89%, 98.88%, and 98.88% on Figshare, and consistent 99.83% on Br35H, indicate its robustness on diverse datasets. We utilized Grad-CAM visualization to provide interpretability by highlighting important regions contributing to classification decisions. These impressive results underscore LiteNet’s capability to accurately detect and classify brain tumors, making it a valuable tool for enhancing diagnostic accuracy in clinical settings, particularly in resource-limited environments.

KW - Classification

KW - EfficientNetB0

KW - Feature fusion

KW - Lightweight feature extraction

KW - MobileNetV2

KW - Transfer learning

UR - https://www.scopus.com/pages/publications/105009303339

U2 - 10.1007/978-981-96-6960-8_10

DO - 10.1007/978-981-96-6960-8_10

M3 - Conference contribution

AN - SCOPUS:105009303339

SN - 9789819669592

T3 - Communications in Computer and Information Science

SP - 132

EP - 147

BT - Neural Information Processing - 31st International Conference, ICONIP 2024, Proceedings

A2 - Mahmud, Mufti

A2 - Doborjeh, Maryam

A2 - Doborjeh, Zohreh

A2 - Wong, Kevin

A2 - Leung, Andrew Chi Sing

A2 - Tanveer, M.

PB - Springer Science and Business Media Deutschland GmbH

Y2 - 2 December 2024 through 6 December 2024

ER -

Nizamani AH, Chen Z, Nizamani AA, Ibrahim AMA. LiteNet: A Lightweight Feature Fusion Model for Brain Tumor Classification. In Mahmud M, Doborjeh M, Doborjeh Z, Wong K, Leung ACS, Tanveer M, editors, Neural Information Processing - 31st International Conference, ICONIP 2024, Proceedings. Springer Science and Business Media Deutschland GmbH. 2025. p. 132-147. (Communications in Computer and Information Science). doi: 10.1007/978-981-96-6960-8_10