Casual3DHDR: High Dynamic Range 3D Gaussian Splatting from casually-captured Videos

Shucheng Gong, Lingzhe Zhao, Wenpu Li, Hong Xie, Yin Zhang, Shiyu Zhao, Peidong Liu · 2025

Read Casual3DHDR: High Dynamic Range 3D Gaussian Splatting from casually-captured Videos by Shucheng Gong, Lingzhe Zhao, Wenpu Li, Hong Xie, Yin Zhang, Shiyu Zhao, Peidong Liu (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-2504-17728.txt.

Description

Rebuilding a 3D scene from video works best with sharp, well-lit frames, but casual phone clips are blurry and unevenly exposed. Shucheng Gong, Lingzhe Zhao, Wenpu Li, Hong Xie, Yin Zhang, Shiyu Zhao, and Peidong Liu tackle that mess head-on. Their method rebuilds wide-range 3D scenes, holding both bright and dark detail, from auto-exposure videos shot on the move.

One model handles the whole chain: the camera's path through time, the shifting exposure, and how the sensor turns light into pixels. All are fitted together in a single stage. On made-up and real datasets the method beat rivals in both toughness and image quality. The team promises to release the code and data.

Catalog record

CREATORShucheng Gong, Lingzhe Zhao, Wenpu Li, Hong Xie, Yin Zhang, Shiyu Zhao, Peidong Liu
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE44 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-17728

Browse the full library · About the Tower of Babel