BCDDM: Branch-Corrected Denoising Diffusion Model for Black Hole Image Generation

Ao liu, Zelin Zhang, Songbai Chen, Cuihong Wen, Jieci Wang · 2025

Read BCDDM: Branch-Corrected Denoising Diffusion Model for Black Hole Image Generation by Ao liu, Zelin Zhang, Songbai Chen, Cuihong Wen, Jieci Wang (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-08528.txt.

Description

Ao Liu and co-authors build a diffusion model that draws black hole images straight from physical settings, in a 2025 paper. The standard way to make such images traces light rays through curved space, which costs heavy computing time.

Their model, BCDDM, adds a branch-correction step and a weighted mix of error measures for stable learning. It trained on about two thousand ray-traced images spanning seven physical settings of the gas flow. Drawn images tracked their settings closely. When the drawn images joined the training pool, a separate network read the settings back far more accurately. The method offers a cheaper path to fitting telescope data of real black holes.

Catalog record

CREATORAo liu, Zelin Zhang, Songbai Chen, Cuihong Wen, Jieci Wang
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE37 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-08528

Browse the full library · About the Tower of Babel