Edge-intelligent Distributed Framework for Brain Tumor Classification Using Hybrid TSK-convolutional Neural Architecture

Read full text Summarise with VIRAT
Authors: Balqees Talal Hasan, Ali M. Ahmed Al-Sabaawi, Hesham Hashim Mohammed, Zaid. J. Al-Araji
Cited by: 0
References: 27

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

The development of tumor classification systems showed an exponential growth in efficiency and accuracy owing to the application of artificial intelligence and medical imaging technology.However, the deployment of these models on resource-limited hardware platforms still poses challenges.In this work, we present LHTSK-Net, an inherently interpretable lightweight hybrid deep learning framework tailored for edge-centric environments with diagnostics focus.The architecture of the LHTSK-Net Network incorporates multi-scale Convolutional neural networks (CNN) combined with a Takagi -Sugeno -Kang (TSK) fuzzy inference framework to provide clinicians with interpretability through an explicit rule-based structure associated with the decisions made by the model.The system operates through a distributed edge-to-server framework which employs Raspberry Pi nodes for its operation while using Redis to handle asynchronous messaging and MySQL to provide centralized data storage.The system achieves a test set classification accuracy of 97.0% while it produces cross-validation results of 95.7% ± 0.005 and maintains an AUC of 1.00 with only 0.198M parameters in a model that has a size of 1.4 MB.The complete edge-toserver architecture operates at an average end-to-end latency of about 455 ms while providing a throughput capacity of 2.2 images per second through its operation on a Raspberry Pi 4. The proposed approach provides an efficient solution with transparent performance capabilities which enables real-time tumor diagnosis in resource-limited environments.

Fields of Study & Subjects

Brain Tumor Detection and Classification IoT and Edge/Fog Computing AI in cancer detection

Publication details

Journal / source
International journal of intelligent engineering and systems
Volume
19
Issue
4
Pages
1156–1170
Publication date
23 March 2026
Language
EN