TY - GEN
T1 - DCGAN-based Medical Image Augmentation to Improve ELM Classification Performance
AU - Rando,
AU - Setiawan, Noor Akhmad
AU - Permanasari, Adhistya Erna
AU - Rulaningtyas, Riries
AU - Suksmono, Andriyan Bayu
AU - Sitanggang, Imas Sukaesih
N1 - Publisher Copyright:
© 2022 IEEE.
PY - 2022
Y1 - 2022
N2 - Cervical cancer is one of the deadliest diseases in women. One of the cervical cancer screening methods is pap smear method. However, using a pap smear method to detect cervical cancer takes a long time for a pathologist to diagnose. Hence, a rapid development of medical computerization for early detection to get the results quickly is needed. This paper proposes synthetic data augmentation by using Deep Convolutional Generative Adversarial Network (DCGAN) to increase number of pap smear samples in dataset. Gray Level Co-occurrence Matrix (GLCM) is employed to extract features from dataset. Classification of 3 classes which are Adenocarcinoma, High-Grade Squamous Intraepithelial Lesion (HSIL), and Squamous Cell Carcinoma (SCC) is conducted using Extreme Learning Machine (ELM). The result shows that the addition of synthetic data improves the performance of ELM with the accuracy of 90%. This accuracy is better than the accuracy of ELM using only the original dataset which is 85%.
AB - Cervical cancer is one of the deadliest diseases in women. One of the cervical cancer screening methods is pap smear method. However, using a pap smear method to detect cervical cancer takes a long time for a pathologist to diagnose. Hence, a rapid development of medical computerization for early detection to get the results quickly is needed. This paper proposes synthetic data augmentation by using Deep Convolutional Generative Adversarial Network (DCGAN) to increase number of pap smear samples in dataset. Gray Level Co-occurrence Matrix (GLCM) is employed to extract features from dataset. Classification of 3 classes which are Adenocarcinoma, High-Grade Squamous Intraepithelial Lesion (HSIL), and Squamous Cell Carcinoma (SCC) is conducted using Extreme Learning Machine (ELM). The result shows that the addition of synthetic data improves the performance of ELM with the accuracy of 90%. This accuracy is better than the accuracy of ELM using only the original dataset which is 85%.
KW - Cervical cancer
KW - DCGAN
KW - ELM
KW - GLCM
KW - Pap Smear
UR - http://www.scopus.com/inward/record.url?scp=85146693352&partnerID=8YFLogxK
U2 - 10.1109/COMNETSAT56033.2022.9994559
DO - 10.1109/COMNETSAT56033.2022.9994559
M3 - Conference contribution
AN - SCOPUS:85146693352
T3 - Proceeding - IEEE International Conference on Communication, Networks and Satellite, COMNETSAT 2022
SP - 206
EP - 211
BT - Proceeding - IEEE International Conference on Communication, Networks and Satellite, COMNETSAT 2022
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 11th IEEE International Conference on Communication, Networks and Satellite, COMNETSAT 2022
Y2 - 3 November 2022 through 5 November 2022
ER -