Underrated detectors EdgePoint2 and CLIDD beat famous ones, says researcher
ducha_aiki · x · 2026-10-30
A computer vision researcher argues that SOTA matchers are the Glue family from @pesarlin and @Parskatt's teams (CVPR/ICCV/ICLR), while the best detectors never appeared at major venues: ALIKED got popular, but EdgePoint2 and CLIDD are amazing yet unknown. A kornia PR with benchmarks is coming soon.
More from Research
- Oxford's LOB-Impact benchmark: generative order book models emerge square-root market impact — oxford-lob · 2026-10-30
- LessWrong: Categorical taboos beat threshold definitions for defining neuralese — LinchZhang · 2026-10-30
- Could a math-peak ASI one-shot AGI? Redditor puts 35% odds within 5 years — Budget_Geologist_574 · 2026-10-30
- Researcher runs experiments entirely with AI agents — adversarial checker agents say results hold — burny_tech · 2026-10-30
- SplatXR: open-source Unity toolkit brings 3D Gaussian splats to Quest 3 and Vision Pro — Scobleizer · 2026-10-30
- Chip-integrated Pockels laser emits 780nm with 1GHz tuning, locked to rubidium D2 line — jwt0625 · 2026-10-30