AI Paper Atlas

Research map / 3D vision / Face and shape

GeoDesc: Learning Local Descriptors by Integrating Geometry Constraints

Zixin Luo, Tianwei Shen, Lei Zhou, Siyu Zhu, Runze Zhang, Yao Yao, Tian Fang, Long Quan

ECCV 2018 · 3D vision › Face and shape · approach: Learned mesh and geometric representations · 195 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
train with crowd-sourced SfM data (ablation)HardNet framework · HPatchesmatching mAP: 52.8 -> 57.0Improved
train with crowd-sourced SfM data (ablation)HardNet framework · HPatchesverification mAP: 88.4 -> 90.1Improved
add proposed batch construction with hardest-in-batch loss (ablation)Config 2 · HPatchesmatching mAP: 57.0 -> 50.2Hurt
add proposed batch construction with hardest-in-batch loss (ablation)Config 2 · HPatchesretrieval mAP: 73.2 -> 70.4Hurt
add modified structured loss E1 to batch construction (ablation)Config 3 · HPatchesmatching mAP: 50.2 -> 58.5Improved
add modified structured loss E1 to batch construction (ablation)Config 3 · HPatchesverification mAP: 89.9 -> 90.9Improved
add geometric loss E2 (ablation)Config 4 · HPatchesmatching mAP: 58.5 -> 59.1Improved
add geometric loss E2 (ablation)Config 4 · HPatchesretrieval mAP: 74.5 -> 74.9Improved
replace distance margin with distance ratio in loss (ablation)structured loss · triplet validation accuracytraining efficiencyImproved
use GeoDesc instead of HardNet descriptorHardNet · Heinly benchmark, Exposurematching score: 34.9 -> 36.3Improved
use GeoDesc instead of HardNet descriptorHardNet · Heinly benchmark, Planarmatching score: 27.4 -> 29.1Improved
use GeoDesc instead of DSP-SIFTDSP-SIFT · Heinly benchmark, JPEG and Blurmatching score: 35.1 vs 34.7 (JPEG), 14.3 vs 14.4 (Blur)No clear effect

Later papers that compared with this method

1 later accepted paper reports GeoDesc as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

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