AI Paper Atlas

Research map / 3D vision / Human pose estimation

SiLK: Simple Learned Keypoints

Pierre Gleize, Weiyao Wang, Matt Feiszli

ICCV 2023 · 3D vision › Human pose estimation · approach: Learned camera pose estimation · 112 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
remove zero padding in convolutions (ablation)SiLK · HPatcheshomography estimation accuracy @1px: 0.59 -> 0.62Improved
remove zero padding in convolutions (ablation)SiLK · ScanNetrotation accuracy @5 deg: 93.7 -> 98.1Improved
remove zero padding in convolutions (ablation)SiLK · ScanNettranslation accuracy @5cm: 75.2 -> 82.9Improved
replace VGGnp-4 by FPN backbone (ablation)SiLK · HPatcheshomography estimation accuracy @1px: 0.62 -> 0.55Hurt
replace VGGnp-4 by UNet backbone (ablation)SiLK · HPatcheshomography estimation accuracy @1px: 0.62 -> 0.41Hurt
shrink backbone to VGGnp-micro (76k params) (ablation)SiLK · HPatcheshomography estimation accuracy @1px: 0.62 -> 0.56; MMA 0.59 -> 0.40Mixed
shrink backbone to VGGnp-micro (ablation)SiLK · HPatchesrepeatability @1px: 0.62 -> 0.64No clear effect
reduce training descriptor map from 146x146 to 82x82 (ablation)SiLK · HPatcheshomography estimation accuracy @1px: 0.59 -> 0.58; training time 5h -> 1.7hNo clear effect
train on ScanNet instead of COCO (ablation)SiLK · HPatchesmean matching accuracy @1px: 0.59 -> 0.48Hurt
train on ImageNet instead of COCO (ablation)SiLK · ScanNettranslation accuracy @5cm: 82.9 -> 83.5No clear effect
increase top-k from 1k to 10k (ablation)SiLK · HPatcheshomography estimation accuracy @1px: 0.53 -> 0.62Improved
use SiLK with simple MNN and no context aggregation vs SuperGlueSiLK · ScanNet relative posepose AUC @5 deg: 16.2 -> 18.0Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy