Channel-wise Noise Scheduled Diffusion for Inverse Rendering in Indoor Scenes

JunYong Choi, Min-cheol Sagong, SeokYeong Lee, Seung-Won Jung, Ig-Jae Kim, Junghyun Cho · 2025

Read Channel-wise Noise Scheduled Diffusion for Inverse Rendering in Indoor Scenes by JunYong Choi, Min-cheol Sagong, SeokYeong Lee, Seung-Won Jung, Ig-Jae Kim, Junghyun Cho (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-2503-09993.txt.

Description

JunYong Choi and five colleagues split a single indoor photo back into geometry, material, and lighting using diffusion models, in a 2025 paper. This inverse rendering problem has no single right answer, so generative methods often offer several plausible splits. Accuracy and variety pull in opposite directions, and earlier models traded one for the other.

Their trick is to schedule noise differently across color channels during training. Two models trained this way split the jobs: one returns a single highly accurate decomposition, the other offers a spread of alternatives. Tests showed gains on both fronts, which carried into practical tasks like inserting objects into a scene and editing materials.

Catalog record

CREATORJunYong Choi, Min-cheol Sagong, SeokYeong Lee, Seung-Won Jung, Ig-Jae Kim, Junghyun Cho
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE45 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-09993

Browse the full library · About the Tower of Babel