CSS-BA: Gated Column Space Bundle Adjustment
zhenjun_zhao · x · 2026-07-20
This paper introduces CSS-BA: a gate-guided column space search method for Bundle Adjustment.
The core idea is to first perform gate-guided column space selection within a geometric prior support set, then construct a low-dimensional subspace using Lanczos/Ritz bases, solve the projected LM system within this subspace, and finally lift the update back to the full camera space. The figure illustrates a workflow comparison between standard LM and CSS-BA.
Experimental results show that on the PhoneSweep dataset, CSS-BA significantly improves pose accuracy compared to Normal-LM and PoBA. Tables further present average improvements across multiple problem families, demonstrating the effectiveness of this constrained subspace search in BA scenarios.
More from Research
- Project APE finds verifier reliability drops when papers contain multiple errors — soumitrashukla9 · 2026-07-22
- Project APE says verifier costs fell about 90x in a year as Chinese open models lead — soumitrashukla9 · 2026-07-22
- OpenAI-linked paper says capability RL can make models more reward-seeking — MariusHobbhahn · 2026-07-22
- Project APE builds its verifier benchmark from 100 AI-written papers with injected errors — soumitrashukla9 · 2026-07-22
- Paper proposes a CRED taxonomy and benchmark to measure research-error detectors — soumitrashukla9 · 2026-07-22
- OpenAI says long-horizon models need safety and alignment checks across full action sequences — rhiever · 2026-07-22