Penerapan Blok SE-NET Pada Deep Learning Inceptionv3 untuk Meningkatkan Deteksi Penyakit Mpox pada Manusia

Penulis

  • M. Bakhara Alief Rachman Sekolah Tinggi Ilmu Manajemen & Ilmu Komputer ESQ, Jakarta
  • Aliyah Kurniasih Sekolah Tinggi Ilmu Manajemen & Ilmu Komputer ESQ, Jakarta
  • Andika Sundawijaya Sekolah Tinggi Ilmu Manajemen & Ilmu Komputer ESQ, Jakarta
  • Ahlijati Nuraminah Sekolah Tinggi Ilmu Manajemen & Ilmu Komputer ESQ, Jakarta

DOI:

https://doi.org/10.25126/jtiik.1077978

Kata Kunci:

Mpox, MSLD (Monkeypox Skin Lesion Dataset), CNN (Convolutional Neural Network), InceptionV3, SE-Net

Abstrak

Mpox atau cacar monyet adalah penyakit yang disebabkan oleh virus monkeypox. Penelitian terdahulu membuktikan sudah tersedia beberapa pre-trained model yang terbukti mampu mendeteksi penyakit mpox dengan menggunakan dataset MSLD (Monkeypox Skin Lesion Dataset) seperti VGG16, ResNet50, InceptionV3, dan penggabungan ketiga model tersebut. Dari penelitian tersebut didapatkan hasil model InceptionV3 memiliki tingkat akurasi paling rendah dengan nilai 74.07% berbanding jauh dengan ResNet50 yang mampu hingga 82.96% dan menjadikannya akurasi tertinggi. Namun, terdapat peluang akurasi model InceptionV3 mampu ditingkatkan. Oleh sebab itu, pada penelitian diimplementasikan arsitektur baru dan penambahan blok SE-Net (Squeeze and Excitation Networks) pada pre-trained model InceptionV3. Untuk training dan evaluasi model akan menggunakan dataset MSLD. Penelitian ini dilaksanakan dengan harapan mampu meningkatkan akurasi pre-trained model InceptionV3 dalam mendeteksi penyakit mpox. Dari hasil penelitian berdasarkan nilai confusion matrix penerapan arsitektur baru berhasil dilakukan terbukti dengan peningkatan akurasi dari 74.07% menjadi 82.22%. Selain itu, penambahan blok SE-Net terhadap arsitektur baru terbukti mampu meningkatkan akurasi menjadi 91.11% dan menjadikan performa InceptionV3 menjadi lebih baik dari akurasi ResNet50. Dari hasil penelitian tersebut memberikan rekomendasi untuk melakukan percobaan dengan mengganti pre-trained model, blok SE-Net, dan jumlah perbandingan dataset antara train, validation, dan test.

 

Abstract

 

Mpox or monkeypox is a disease caused by the monkeypox virus. Previous research has proven that there are several pre-trained models that are proven to be able to detect mpox disease using MSLD (Monkeypox Skin Lesion Dataset) datasets such as VGG16, ResNet50, InceptionV3, and a combination of these three models. From this research, it was found that the InceptionV3 model has the lowest level of accuracy with a value of 74.07% compared to ResNet50 which is capable of up to 82.96% and makes it the highest accuracy. However, there is a chance that accuracy can be improved. Therefore, this research will apply a new architecture and SE-Net blocks to the InceptionV3 pre-trained model using the MSLD dataset. From the results of research based on the value of the confusion matrix the application of the new architecture was successfully carried out as evidenced by an increase in accuracy from 74.07% to 82.22%. In addition, the addition of the SE-Net block to the new architecture is proven to be able to increase accuracy to 91.11% and make InceptionV3's performance better than ResNet50's accuracy. The results of this study provide recommendations for conducting experiments by changing the pre-trained model, the SE-Net block, and the number of dataset comparisons between train, validation, and test.

Downloads

Download data is not yet available.

Referensi

ALI, S.N., AHMED, MD.T., PAUL, J., JAHAN, T., SANI, S.M.S., NOOR N. and HASAN, T., 2022. Monkeypox Skin Lesion Detection Using Deep Learning Models: A Feasibility Study. [online] Available at: <http://arxiv.org/abs/2207.03342>.

BAJAJ, V. and SINHA, G.R., 2022. Artificial Intelligence-Based Brain-Computer Interface. Elsevier. https://doi.org/10.1016/C2021-0-00055-9.

GOOGLE, 2023. Advanced Guide to Inception v3. [online] Available at: <https://cloud.google.com/tpu/docs/inception-v3-advanced> [Accessed 9 October 2023].

HU, J., SHEN, L. and SUN, G., 2018. Squeeze-and-Excitation Networks. [online] Available at: <http://image-net.org/challenges/LSVRC/2017/results>.

KAGGLE, 2023. Monkeypox Skin Lesion Dataset. [online] 2022. Available at: <https://www.kaggle.com/datasets/nafin59/monkeypox-skin-lesion-dataset> [Accessed 9 October 2023].

KARHUMEN, J., RAIKO, T. and CHO, K., 2015. Unsupervised deep learning. In: Advances in Independent Component Analysis and Learning Machines. Elsevier. pp.125–142. https://doi.org/10.1016/B978-0-12-802806-3.00007-5.

El NAQA, I. and MURPHY, M.J., 2015. What Is Machine Learning? In: Machine Learning in Radiation Oncology. Cham: Springer International Publishing. pp.3–11. https://doi.org/10.1007/978-3-319-18305-3_1.

PARKER, S. and BULLER, R.M., 2013. A review of experimental and natural infections of animals with monkeypox virus between 1958 and 2012. Future Virology, 8(2), pp.129–157. https://doi.org/10.2217/fvl.12.130.

SARKER, I.H., 2021. Deep Learning: A Comprehensive Overview on Techniques, Taxonomy, Applications and Research Directions. SN Computer Science, 2(6), p.420. https://doi.org/10.1007/s42979-021-00815-1.

SZEGEDY, C., VANHOUCKE, V., IOFFE, S., SHLENS, J. and WOJNA, Z., 2016. Rethinking the Inception Architecture for Computer Vision. In: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). IEEE. pp.2818–2826. https://doi.org/10.1109/CVPR.2016.308.

WHO, 2022a. Monkeypox Fact Sheets. [online] Available at: <https://www.who.int/news-room/fact-sheets/detail/monkeypox> [Accessed 22 December 2023].

WHO, 2022b. Monkeypox Response in Europe. [online] Available at: <https://www.who.int/news-room/feature-stories/detail/monkeypox-response-in-europe> [Accessed 18 December 2023].

WHO, 2022c. Mpox (monkeypox). [online] Available at: <https://www.who.int/news-room/questions-and-answers/item/monkeypox> [Accessed 10 December 2022].

YE, J. and TOYAMA, N., 2021. Benchmarking Deep Learning Models for Automatic Ultrasonic Imaging Inspection. IEEE Access, 9, pp.36986–36994. https://doi.org/10.1109/ACCESS.2021.3062860.

ZUMLA, A., VALDOLEIROS, S.R., HAIDER, N., ASOGUN, D., NTOUMI, F., PETERSEN, E. and KOCK, R., 2022. Monkeypox outbreaks outside endemic regions: scientific and social priorities. The Lancet Infectious Diseases, 22(7), pp.929–931. https://doi.org/10.1016/S1473-3099(22)00354-1.

Diterbitkan

30-12-2023

Cara Mengutip

Penerapan Blok SE-NET Pada Deep Learning Inceptionv3 untuk Meningkatkan Deteksi Penyakit Mpox pada Manusia. (2023). Jurnal Teknologi Informasi Dan Ilmu Komputer, 10(7), 1447-1452. https://doi.org/10.25126/jtiik.1077978