Version 1
: Received: 26 October 2024 / Approved: 28 October 2024 / Online: 28 October 2024 (14:54:23 CET)
How to cite:
Diyasa, I. G. S. M.; Humairah, S.; Puspaningrum, E. Y.; Durry, F. D.; Lestari, W. D.; Caesarendra, W. Data Augmentation of Breast Ultrasound Images Using Wasserstein Generative Adversarial Networks. Preprints2024, 2024102205. https://doi.org/10.20944/preprints202410.2205.v1
Diyasa, I. G. S. M.; Humairah, S.; Puspaningrum, E. Y.; Durry, F. D.; Lestari, W. D.; Caesarendra, W. Data Augmentation of Breast Ultrasound Images Using Wasserstein Generative Adversarial Networks. Preprints 2024, 2024102205. https://doi.org/10.20944/preprints202410.2205.v1
Diyasa, I. G. S. M.; Humairah, S.; Puspaningrum, E. Y.; Durry, F. D.; Lestari, W. D.; Caesarendra, W. Data Augmentation of Breast Ultrasound Images Using Wasserstein Generative Adversarial Networks. Preprints2024, 2024102205. https://doi.org/10.20944/preprints202410.2205.v1
APA Style
Diyasa, I. G. S. M., Humairah, S., Puspaningrum, E. Y., Durry, F. D., Lestari, W. D., & Caesarendra, W. (2024). Data Augmentation of Breast Ultrasound Images Using Wasserstein Generative Adversarial Networks. Preprints. https://doi.org/10.20944/preprints202410.2205.v1
Chicago/Turabian Style
Diyasa, I. G. S. M., Wahyu Dwi Lestari and Wahyu Caesarendra. 2024 "Data Augmentation of Breast Ultrasound Images Using Wasserstein Generative Adversarial Networks" Preprints. https://doi.org/10.20944/preprints202410.2205.v1
Abstract
Breast cancer is among the most prevalent cancers in Indonesia. One of the methods for early detection of breast cancer is utilizing ultrasound images to classify breast conditions into normal, benign tumors, or cancer. The advent of deep learning technology facilitates the image analysis process, such as the use of Transfer Learning Convolutional Neural Networks (CNNs). Generally, CNN models require large and balanced datasets to perform well in classification tasks. However, medical datasets like breast ultrasound images tend to be limited and imbalanced. The Wasserstein GAN (WGAN) is a generative data augmentation method capable of producing synthetic images by learning the pattern of the distribution of real image data. The implementation of the Wasserstein distance results in the training process of WGAN demonstrating stability from epochs 3000, 2500, and 3000 out of a total of 5000 epochs. The quality of the synthetic images improves with an increasing number of training iterations. By using WGAN, all of the evaluation metrics of each classifier are increasing, with the best accuracy score is achieved by VGG16 model with 83.33% accuracy
Keywords
Wasserstein GAN; Breast Cancer; Ultrasound Image; Augmentation
Subject
Computer Science and Mathematics, Artificial Intelligence and Machine Learning
Copyright:
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
