Article Per Year (5 Year)
p-Index From 2021 - 2026
0.23
P-Index
This Author published in this journals
All Journal Sinkron : Jurnal dan Penelitian Teknik Informatika
Muh., Satriawan
Unknown Affiliation
Computer Science & IT

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Improving Brain Tumor Classification Performance Using EfficientNetB0 Integrated with CBAM Attention Mechanism At Taqwa, Abd Salam; Muh., Satriawan; Sheila Eunike, Kakisina; Puput, Kusuma Dewi; Fettyana; Ayutri, Wahyuni
Sinkron : jurnal dan penelitian teknik informatika Vol. 10 No. 2 (2026): Article Research April, 2026
Publisher : Politeknik Ganesha Medan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33395/sinkron.v10i2.15931

Abstract

Accurate classification of brain tumors using magnetic resonance imaging (MRI) requires robust automated methods to support clinical diagnosis, particularly when tumor types present subtle visual distinctions. In this study, the Convolutional Block Attention Module (CBAM) is incorporated into the EfficientNetB0 architecture to improve feature representation for multi-class brain tumor classification. The performance of the proposed model is evaluated against the baseline EfficientNetB0 under identical training and testing conditions. EfficientNetB0 with CBAM achieves a training accuracy of 99.76% and a validation accuracy of 99.45%, with corresponding training and validation losses of 0.0085 and 0.0241. On an independent test dataset, the model attains a test accuracy of 99.25% and a loss of 0.0207. In comparison, the baseline EfficientNetB0 model attains a training accuracy of 52.68%, validation accuracy of 46.20%, and test accuracy of 43.32%, accompanied by significantly higher loss values. At the class level, the proposed model demonstrates macro-average precision, recall, and F1-score of 0.99, whereas the baseline model yields macro-average values of approximately 0.54 for precision and recall, and 0.50 for F1-score. Although CBAM integration increases computational time per evaluation step from 395 ms to 601 ms, the marked improvement in classification accuracy and error reduction underscores the value of attention mechanisms. These results demonstrate that attention-based feature refinement significantly enhances deep learning performance for medical image classification, particularly in multi-class brain tumor diagnosis.
Co-Authors At Taqwa, Abd Salam Ayutri, Wahyuni Fettyana Puput, Kusuma Dewi Sheila Eunike, Kakisina