Perbandingan EfficientNet dan MobileNet dalam Klasifikasi Kanker Kulit Menggunakan Explainable Artificial Intelligence
DOI:
https://doi.org/10.52436/1.jpti.1748Kata Kunci:
Class Weighting, Deep Learning, EfficientNet-B4, Grad-CAM, Kanker Kulit, MobileNetV2Abstrak
Deteksi kanker kulit yang akurat dan dini sangat penting untuk meningkatkan tingkat kelangsungan hidup pasien secara global. Namun, meskipun Deep Convolutional Neural Networks telah menunjukkan kinerja luar biasa dalam klasifikasi dermatologis, sifat black-box yang melekat sering membatasi adopsi klinis dan kepercayaan di antara para profesional medis. Untuk mengatasi masalah ini, penelitian ini bertujuan membandingkan dua arsitektur jaringan terkemuka, EfficientNet-B4 dan MobileNetV2, untuk mengklasifikasikan lesi kulit guna menemukan model yang paling dapat diinterpretasikan. Metode penelitian yang diterapkan menggunakan dataset HAM10000 yang sangat tidak seimbang, terdiri dari tujuh kelas lesi yang berbeda. Untuk memitigasi ketidakseimbangan kelas tanpa mendistorsi citra medis asli melalui teknik oversampling sintetis, strategi pembobotan kelas (class weighting) yang presisi diimplementasikan selama fase pelatihan. Selanjutnya, kedua model melalui tahap fine-tuning dan dievaluasi menggunakan metrik kuantitatif yang ketat serta penilaian visual kualitatif. Hasil penelitian menunjukkan bahwa EfficientNet-B4 secara konsisten mengungguli MobileNetV2 di seluruh dimensi statistik. EfficientNet-B4 mencapai akurasi pengujian sebesar 80,44% dibandingkan dengan 79,34% untuk MobileNetV2, bersama dengan Macro F1-Score yang jauh lebih unggul (0,75 berbanding 0,68). Hal ini menyoroti ketangguhan EfficientNet-B4 dalam mengklasifikasikan kelas minoritas secara akurat. Selain itu, Gradient-weighted Class Activation Mapping diintegrasikan untuk memberikan penjelasan visual di semua kelas. Temuan mengungkapkan bahwa EfficientNet-B4 memberikan lokalisasi area lesi kritis yang lebih tajam, terutama pada kasus parah seperti Melanoma, dibandingkan dengan fokus MobileNetV2 yang lebih menyebar. Kesimpulannya, penelitian ini menetapkan EfficientNet-B4 sebagai arsitektur yang secara fundamental lebih unggul dan sangat dapat diinterpretasikan untuk lingkungan dermatologis klinis, menawarkan akurasi tinggi, stabilitas kelas minoritas, dan penjelasan visual yang andal.
Unduhan
Referensi
M. Denny Richardo and H. Satria, "Comparative Analysis of Explainable AI Models for Pneumonia Detection in Chest X-rays Using Grad-CAM," Jurnal Sisfokom (Sistem Informasi dan Komputer), vol. 14, no. 4, pp. 475–483, 2025, https://doi.org/10.32736/sisfokom.v14i4.2450
Khan, M., Kurada, S. B., Menshawi, A. M., Malik, V., & Altaf, M. M. (2026). Novel bioluminescent oceanic optimization with explainable AI for accurate skin cancer diagnosis and enhanced dermatologist support. Journal of King Saud University - Science, 38(2). https://doi.org/10.25259/JKSUS_1183_2025
Cherukuri, S., & Vara Prasad, S. D. (2025). EnsembleSkinNet: a transfer learning-based framework for efficient skin cancer detection with explainable AI integration. Frontiers in Oncology, 15. https://doi.org/10.3389/fonc.2025.1699960
S. M. R. Swapno et al., "Towards unbiased skin cancer classification using deep feature fusion," Scientific Reports, vol. 14, no. 1, p. 5432, 2024. https://doi.org/10.1186/s12911-025-02889-w
F. Mahmud, M. M. Mahfiz, M. Z. I. Kabir and Y. Abdullah, "An Interpretable Deep Learning Approach for Skin Cancer Categorization," 2023 26th International Conference on Computer and Information Technology (ICCIT), Cox's Bazar, Bangladesh, 2023, pp. 1-6, https://doi.org/10.1109/ICCIT60459.2023.10508527
A. H. Shahin et al., "Skin cancer classification using EfficientNet architecture," Bulletin of Electrical Engineering and Informatics, vol. 13, no. 4, pp. 2716-2728, 2024. https://doi.org/10.11591/eei.v13i4.7159
Hamim, S. A., Tamim, M. U. I., Mridha, M. F., Safran, M., & Che, D. (2025). SmartSkin-XAI: An Interpretable Deep Learning Approach for Enhanced Skin Cancer Diagnosis in Smart Healthcare. Diagnostics, 15(1). https://doi.org/10.3390/diagnostics15010064
Shahadha, R. S., & Al-Khateeb, B. (2025). Dual Convolutional Neural Network for Skin Cancer Classification. Journal of Cybersecurity and Information Management, 15(2), 35–42. https://doi.org/10.54216/JCIM.150204
Alsaidi, M., Jan, M.T., Altaher, A. et al. Tackling the class imbalanced dermoscopic image classification using data augmentation and GAN. Multimed Tools Appl 83, 49121–49147 (2024). https://doi.org/10.1007/s11042-023-17067-1
Mridha, K., Uddin, M. M., Shin, J., Khadka, S., & Mridha, M. F. (2023). An Interpretable Skin Cancer Classification Using Optimized Convolutional Neural Network for a Smart Healthcare System. IEEE Access, 11, 41003–41018. https://doi.org/10.1109/ACCESS.2023.3269694
Thwin, S. M., Park, H. S., & Seo, S. H. (2025). A Trustworthy Framework for Skin Cancer Detection Using a CNN with a Modified Attention Mechanism. Applied Sciences (Switzerland), 15(3). https://doi.org/10.3390/app15031067
Hidayat, T., Khasanah, N., Firasari, E., Kurniawati, L., & Hermaliani, E. H. (n.d.). Comparative Evaluation of CNN Architectures for Skin Cancer Classification. In IJACSA) International Journal of Advanced Computer Science and Applications. 16.10 (2025). http://dx.doi.org/10.14569/IJACSA.2025.0161007
Himel, G. M. S., Islam, M. M., Al-Aff, K. A., Karim, S. I., & Sikder, M. K. U. (2024). Skin Cancer Segmentation and Classification Using Vision Transformer for Automatic Analysis in Dermatoscopy-Based Noninvasive Digital System. International Journal of Biomedical Imaging, 2024. https://doi.org/10.1155/2024/3022192
Fiaz, M., Shoaib Khan, M. B., Khan, A. H., Bilal, A., Abdullah, M., Darem, A. A., & Sarwar, R. (2025). An explainable hybrid deep learning framework for precise skin lesion segmentation and multi-class classification. Frontiers in Medicine, 12. https://doi.org/10.3389/fmed.2025.1681542
Esteva, A., Chou, K., Yeung, S., Naik, N., Madani, A., Mottaghi, A., Liu, Y., Topol, E., Dean, J., & Socher, R. (2021). Deep learning-enabled medical computer vision. In npj Digital Medicine (Vol. 4, Issue 1). Nature Research. https://doi.org/10.1038/s41746-020-00376-2
Y. Arun and G. S. Viknesh, "Leaf Classification for Plant Recognition Using EfficientNet Architecture," 2022 IEEE Fourth International Conference on Advances in Electronics, Computers and Communications (ICAECC), Bengaluru, India, 2022, pp. 1-5, doi: https://doi.org/10.1109/icaecc54045.2022.9716637
Khiat, S., Mahmoudi, S. A., Stassin, S., Boukerroui, L., Senaï, B., & Mahmoudi, S. (2025). An Efficient Explainability of Deep Models on Medical Images. Algorithms, 18(4). https://doi.org/10.3390/a18040210
Aras, H.H., Do?an, N. A comparative analysis for skin cancer detection by using explainable deep learning. Neural Comput & Applic 38, 50 (2026). https://doi.org/10.1007/s00521-025-11809-y
Ramadhan, M. E., & Zeniarja, J. (n.d.). Sistemasi: Jurnal Sistem Informasi Implementasi Deep Transfer Learning dan Explainable AI dalam Klasifikasi Kanker Kulit Implementation of Deep Transfer Learning and Explainable AI in Skin Cancer Classification. Vol 14. No 5 (2025). https://doi.org/10.32520/stmsi.v14i5.5425
Roy, A.L., Al Maruf, A., Sony, R.I. et al. Fostering trust and interpretability in skin cancer classification: a hybrid framework of deep learning and machine learning with explainable AI. J Ambient Intell Human Comput 16, 879–894 (2025). https://doi.org/10.1007/s12652-025-04984-2
S. Nurmaini, and H. Rahmat, “Deep Learning for Improving the Effectiveness of Routine Prenatal Screening for Major Congenital Heart Diseases.” International Journal of Medical Imaging, vol. 9, no. 2, 2023, doi: https://doi.org/10.1155/2023/8342974
Aboulmira A, Hrimech H, Lachgar M, Hanine M, Garcia CO, Mezquita GM, Ashraf I. Hybrid Model with Wavelet Decomposition and EfficientNet for Accurate Skin Cancer Classification. J Cancer. 2025 Jan 1;16(2):506-520. doi: 10.7150/jca.101574
Behara, K., Behara, S. (2025). Explainable Artificial Intelligence in Dermatology: Overview of Data Scarcity for Skin Cancer Detection in Black Skin. In: Naidoo, P., Sibanda, M. (eds) Measurement, Automation, and Control in Artificial Intelligence and Machine Learning. Signals and Communication Technology. Springer, Cham. https://doi.org/10.1007/978-3-031-91802-5_12
Liu, X., Sangers, T. E., Nijsten, T., Kayser, M., Pardo, L. M., Wolvius, E. B., Roshchupkin, G. v., & Wakkee, M. (2024). Predicting skin cancer risk from facial images with an explainable artificial intelligence (XAI) based approach: a proof-of-concept study. eClinicalMedicine, 71. https://doi.org/10.1016/j.eclinm.2024.102550
Dagnaw, G. H., el Mouhtadi, M., & Mustapha, M. (2024). Skin cancer classification using vision transformers and explainable artificial intelligence. Journal of Medical Artificial Intelligence, 7(June). https://doi.org/10.21037/jmai-24-6
