Perbandingan Arsitektur ResNet50 dan ResNet101 dalam Klasifikasi Kanker Serviks pada Citra Pap Smear

Za'imatun Niswati; Rahayuning Hardatin; Meia Noer Muslimah; Siti Nur Hasanah

doi:10.30998/faktorexacta.v14i3.10010

Perbandingan Arsitektur ResNet50 dan ResNet101 dalam Klasifikasi Kanker Serviks pada Citra Pap Smear

Za'imatun Niswati^(1*), Rahayuning Hardatin⁽²⁾, Meia Noer Muslimah⁽³⁾, Siti Nur Hasanah⁽⁴⁾

(1) (Scopus ID:57205060572) Universitas Indraprasta PGRI
(2) IPB University
(3) IPB University
(4) IPB University
(*) Corresponding Author

Abstract

Cervical cancer is one of the most deadly types of cancer in women. According to Ferlay et al. (2018) cervical cancer ranks second for the type of cancer that attacks women the most. Data from the Indonesian Ministry of Health, there are at least 15000 cases of cervical cancer every year in Indonesia. This cancer is a type of tumor that develops in the epithelial tissue of the cervix. In addition to HPV vaccination, cervical cancer detection can also be carried out with a Pap smear test and VIA examination supported by medical image tests such as CT scan, microscopic and MRI (Akbar et al. 2021). Pap smear test is a type of test to detect cervical cancer which is quite widely used because the cost of the test is cheaper than the HPV vaccination. This test is carried out by taking samples of uterine cells which are then analyzed for early detection of cervical cancer (BPJS Kesehatan 2020). Through a pap smear can be found the presence of HPV infection and abnormal cells that can turn into cancer cells. The purpose of this research is to apply the ResNet50 and ResNet101 architectures on pap smear images to identify cervical cancer and evaluate the performance of the ResNet50 and ResNet101 architectures in the classification of cervical cancer on pap smear images. In this study, CNN ResNet50 and ResNet101 were used to classify cervical cancer on pap smear images. This study has created two models to predict the grade of cervical cancer on pap smear images. The ResNet50 architecture gets 91% accuracy while the ResNet101 architecture gets 89%. Although the architecture of ResNet101 is more complex than ResNet50, but if viewed from the results of the model evaluation, ResNet101 has a worse performance. This is due to the relatively small training data when trained with a large architecture such as ResNet101, not necessarily resulting in better accuracy.

Full Text:

PDF (Indonesian)

References

Ferlay J, Ervik M, Lam F, Colombet M, Mery L, Piñeros M, Znaor A, Soerjomataram I and F., “Global Cancer Observatory: Cancer Today.,” International Agency for Research on Cancer. 2018.

Rahaman Md, Li C, Wu X, Yao Y, Hu Z, Jiang T, Li X, “A Survey for Cervical Cytopathology Image Analysis Using Deep Learning.,” IEEE Acces, 2020, doi:10.1109/ACCESS.2020.2983186.

Akbar H, Sandfreni. “Klasifikasi Kanker Serviks Menggunakan Model Convolutional Neural Network (Alexnet).,” J. Inform. dan Komputer., 2021, doi: 10.33387/jiko.

Badan Penyelenggara Jaminan Sosial Kesehatan Republik Indonesia. 2021.( https://bpjs-kesehatan.go.id/bpjs/index.php/post/read/2014/262/Jangan-Khawatir-BPJS-Kesehatan-Menjamin-Deteksi-Sebelum-Kanker-Serviks-Menyerang/berita)

Kurniawan R, Sasmito Kartikaning DE, Suryani F “Klasifikasi Sel Serviks Menggunakan Analisis Fitur Nuclei pada Citra Pap Smear.,” 2013. Seminar Nasional Informatika Medis (SNIMed) IV.

Plissiti M, Dimitrakopoulos P, Sfikas G, Nikou C, Krikoni O, “Sipakmed: A New Dataset for Feature and Image Based Classification of Normal and Pathological Cervical Cells in Pap Smear Images.,” pp. 3144-3148., 2018, doi: 10.1109/ICIP.2018.8451588.

Q. Liao, Y. Ding, Z. L. Jiang, X. Wang, C. Zhang, and Q. Zhang, “Multi-task deep convolutional neural network for cancer diagnosis,” Neurocomputing, vol. 348, pp. 66–73, 2019, doi: 10.1016/j.neucom.2018.06.084.

Agrawal H, Kalantri M, “Comparative Analysis of different Convolutional NeuralNetwork Algorithm for Image Classification.,” ISRAJET., vol. 8(9)., 2020.

Suyanto, Machine Learning Tingkat Dasar dan Lanjut. Bandung: Penerbit Informatika., 2018.

“Convolutional neural network". https://en.wikipedia.org/wiki/Convolutional_neural_network

He, K., Zhang, X., Ren, S., & Sun, “Deep Residual Learning for Image Recognition Kaiming.,” IIEE, vol. 32(5), 428, 2015, https://doi.org/10.1246/cl.2003.428.

M. Soh, “Learning CNN-LSTM Architectures for Image Caption Generation. Nips, (c), 1–9.,” 2016. https://cs224d.stanford.edu/reports/msoh.pdf.

P. Arnav, A., Jang, H., & Maloo, Image Captioning Using Deep Learning. 2017.

A. Primawati, I. Mutia, “Analisis Klasifikasi Populasi Ternak Kambing dan Domba dengan Model Convolutional Neural Network,” Fakt. Exacta, vol. 14, no. 1, pp. 22–33, 2021, [Online]. Available: http://journal.lppmunindra.ac.id/index.php/Faktor_Exacta.

Aneja, J., Deshpande, A., & Schwing, “Convolutional Image Captioning. Proceedings,” 2018, https://doi.org/10.1109/CVPR.2018.00583.

“ResNet." https://pytorch.org/hub/pytorch_vision_resnet/.

S. Das, “CNN Architectures: LeNet, AlexNet, VGG, GoogLeNet, ResNet and more,” [Online]. Available: https://medium.com/analytics-vidhya/cnns-architectures-lenet-alexnet-vgg-googlenet-resnet-and-more-666091488df5.

O'Shea K, Nash R. An Introduction to Convolutional Neural Networks. ArXiv e-prints., 2015.

Arel I, Rose D C, “Deep machine learning-a new frontier in artificial intelligence research [research frontier].,” IEEE Comput. Intell. Mag., vol. 5(4): 13–1, 2010.

“SIPaKMeD Database,” [Online]. Available: https://www.cs.uoi.gr/~marina/sipakmed.html.

DOI: http://dx.doi.org/10.30998/faktorexacta.v14i3.10010