HETA++: A Novel Global SfM Pipeline
zhenjun_zhao · x · 2026-07-20
HETA++ introduces a novel pipeline for global SfM: starting with view track graph construction, rotation averaging, and feature track building, followed by relative translation refinement, global filtering, and hybrid explicit initialization, and finally completing the reconstruction via selective angular refinement, selective Bundle Adjustment, and full BA.
The paper presents camera pose AUC and reconstruction time comparisons on the 1DSfM dataset, showing that HETA++ achieves a better accuracy/speed trade-off across multiple scenes. Tables also display comparisons with methods like COLMAP, CReTA, LiGT, HETA, and GLOMAP.
More from Research
- Stanford Team Introduces Gigatoken, the World's Fastest Tokenizer — StanfordAILab · 2026-07-22
- Tabul AI launches Metal TreeSHAP to speed up Shapley values on Apple silicon — Scobleizer · 2026-07-22
- Reddit points to OpenAI’s ChatGPT Ads page — EcstaticAsparagus509 · 2026-07-22
- Open-source runtime lets each repo define its own AI code reviewer — ibabufrik · 2026-07-22
- DeepSWE: A New Benchmark for Evaluating AI Coding Agents on Real GitHub Issues — pmz · 2026-07-22
- A Rust space-economy sim runs hundreds of autonomous ships, built with Claude — kalcode · 2026-07-22