Sparse-view Pose Estimation and Reconstruction via Analysis by Generative Synthesis
Qitao Zhao Shubham Tulsiani · 2024
Read Sparse-view Pose Estimation and Reconstruction via Analysis by Generative Synthesis by Qitao Zhao Shubham Tulsiani (2024) 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-2412-03570.txt.
Description
Qitao Zhao and Shubham Tulsiani tackle building three-dimensional models from just a few photos, in a paper from 2024. Good models need exact camera poses, yet good poses need a model, so the two tasks lean on each other. With few views, evidence is thin and pose errors can wreck the result.
Their method adds learned priors that imagine unseen views, which guides the model toward plausible shapes. It also hunts for outlier poses and fixes them with a mix of broad search and fine tuning. On real and synthetic data sets, it sharpened pose accuracy for several base systems and produced cleaner models than current baselines.
Catalog record
| CREATOR | Qitao Zhao Shubham Tulsiani |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 44 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-03570