Generating View-Specific Chest X-rays Using a Vision-Language Model

A novel approach called ViewXGen that can generate synthetic chest X-rays with specific views, such as posterior-anterior (PA), anterior-posterior (AP), and lateral, by leveraging multi-view chest X-rays and introducing specialized view-specific tokens.