Diffusion pruning for rapidly and robustly selecting global correspondences using local isometry
- Submitting institution
-
Swansea University / Prifysgol Abertawe
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 20809
- Type
- D - Journal article
- DOI
-
10.1145/2517967
- Title of journal
- ACM Transactions on Graphics
- Article number
- -
- First page
- 1
- Volume
- 33
- Issue
- 1
- ISSN
- 1557-7368
- Open access status
- Out of scope for open access requirements
- Month of publication
- December
- Year of publication
- 2014
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
-
- Research group(s)
-
-
- Citation count
- 16
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Surface registration underpins many graphics/vision applications; yet the core to establish fast and reliable correspondences is always a challenge for real-life scans with complex deformations. Previous work compromised with (near-)isometry. We innovated the "locally isometric globally consistent" (one of the few non-isometric) model that better fits real-life deformation. The technique handles surfaces with noises, holes, topological issues; was evaluated on 5 different data sources, and shown 300x faster, 2x more accurate, 100x memory efficient. It is versatile, lays the foundation for 5+ new registration/correspondence/matching techniques; presented in SIGGRAPH2014; led to an ICCM2019 distinguished paper award for a follow-up SGP(CGF) paper.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -