Quality analysis of DCGAN-generated mammography lesions

TL;DR

This study evaluates the realism and feature distribution of mammography lesions generated by DCGANs, showing that synthetic images closely resemble real ones and are often indistinguishable by expert radiologists.

Contribution

It provides a comprehensive analysis of the quality of DCGAN-generated mammogram lesions, including visual, feature space, and observer-based evaluations.

Findings

Generated images have similar feature distribution to real images.

Radiologists often cannot distinguish synthetic from real lesions.

Synthetic images may enhance data diversity for medical imaging applications.

Abstract

Medical image synthesis has gained a great focus recently, especially after the introduction of Generative Adversarial Networks (GANs). GANs have been used widely to provide anatomically-plausible and diverse samples for augmentation and other applications, including segmentation and super resolution. In our previous work, Deep Convolutional GANs were used to generate synthetic mammogram lesions, masses mainly, that could enhance the classification performance in imbalanced datasets. In this new work, a deeper investigation was carried out to explore other aspects of the generated images evaluation, i.e., realism, feature space distribution, and observers studies. t-Stochastic Neighbor Embedding (t-SNE) was used to reduce the dimensionality of real and fake images to enable 2D visualisations. Additionally, two expert radiologists performed a realism-evaluation study. Visualisations showed that the generated images have a similar feature distribution of the real ones, avoiding outliers. Moreover, Receiver Operating Characteristic (ROC) curve showed that the radiologists could not, in many cases, distinguish between synthetic and real lesions, giving 48% and 61% accuracies in a balanced sample set.