What question did this study set out to answer?

The aim is to improve GAN performance in generating synthetic breast ultrasound images for enhanced dataset augmentation.

February 14, 2026Open Access

Augmentation of Small Ultrasound Databases: A Practical Approach

Key Points

The aim is to improve GAN performance in generating synthetic breast ultrasound images for enhanced dataset augmentation.
Introduced Enhanced Deep Convolutional GAN design with modified Discriminator and Generator.
Implemented spectral normalization, Squeeze-and-Excitation blocks, and Scaled Exponential Linear Unit activation.
Established stopping criteria based on loss function convergence and Discriminator constraints.
Utilized visual evaluation by radiologists and image processing metrics for postprocessing.
Achieved a 5% performance improvement over non-augmented data using Deep Learning classification with ResNet-18.
Outperformed previous DCGAN-based augmentation by 3%.
Model achieved high scores on Inception Score, Structural Similarity Index, and Mean Squared Error metrics.

Abstract

Generative Adversarial Networks (GANs) have emerged as a promising tool for augmenting medical image datasets used by AI solutions. However, GANs trained on small datasets (300–500 images) frequently encounter mode collapse, overfitting, and instability, which hinder their practical application. Many GAN-generated images look unrealistic. The Enhanced Deep Convolutional GAN (EDCGAN) is introduced to generate high-quality synthetic images of breast US (BUS). The model includes an experimental design for the Discriminator and Generator. The main components are spectral normalization (SN), the Squeeze-and-Excitation (SE) block, and the Scaled Exponential Linear Unit (SELU). One of the basic versions of DCGAN is considered for the proposed modifications. The stopping criteria are based on the convergence of the smoothed loss function and the constraints imposed on the Discriminator. The contribution is a combination of the above modifications and postprocessing based on the visual evaluation by radiologists and selected image processing metrics. The Inception Score (IS), the Structural Similarity Index (SSIM), and the Mean Squared Error (MSE) comply with the results obtained in the preceding works. The efficiency of augmenting the US data has been verified on a DL classification based on ResNet-18. The tests against training on a non-augmented data outperform ResNet by 5% and by the data augmented by the previous DCGAN by 3%. These numbers are substantial since this variant of ResNet has been pre-trained on 1000 categories by ImageNet-1K, including 1.28 million images. Additionally, the model wins the “Guess-the-real-image” game, competing with seven preceding GANs.

Bookmark

View Full Paper