Request for adding CheXGenBench: A Unified Benchmark For Fidelity, Privacy and Utility of Synthetic Chest Radiographs #24
Description
Request for adding the following paper - CheXGenBench: A Unified Benchmark For Fidelity, Privacy and Utility of Synthetic Chest Radiographs.
The paper presents the first benchmark for evaluating text-to-image models for synthetic chest X-ray generation. It also simultaneously assesses each T2I model in terms of (1) Generative Fidelity, (2) Privacy Leakge, (3) Utility for Downstream Tasks. The paper evaluated frontier T2I models on the MIMIC-CXR dataset and also releases the SoTA model for synthetic chest X-ray generation.
Project Page: https://raman1121.github.io/CheXGenBench/
GitHub: https://github.com/Raman1121/CheXGenBench.