Accepted to GAISynMeD @ ACCV 2024
Authors: William Cagas, Chan Ko, Blake Hsiao, Shryuk Grandhi, Rishi Bhattacharya, Kevin Zhu, Michael Lam
The proliferation of machine learning models in diverse clinical applications has led to a growing need for high-fidelity medical image training data. This work experimentally establishes benchmarks that measure the relationship between a sample dataset size and the fidelity of the generated images, given the dataset's distribution of image complexities. We analyze statistical metrics based on delentropy, an image complexity measure rooted in Shannon's entropy. Across both StyleGAN 3 and SPADE-GAN, general performance improved with increasing training set size but suffered with increasing complexity.
