99,349 frames

Emerging

1papers using it

2026first seen

The '99,349 frames' dataset contains a large collection of digital subtraction angiography (DSA) images used to train a semantically conditioned latent diffusion model for synthesizing arterial-phase DSA frames while evaluating their clinical realism through expert ratings.

🔎 Find this dataset

Papers using 99,349 frames (1)

Semantically Conditioned Diffusion Models for Cerebral DSA Synthesis2026