AI Paper Atlas

Research map / 3D vision / Face and shape

The Power of Points for Modeling Humans in Clothing

Qianli Ma, Jinlong Yang, Siyu Tang, Michael J. Black

ICCV 2021 · 3D vision › Face and shape · approach: Clothed human modeling · 101 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 one cross-outfit model on all outfits instead of per-outfit models (ablation)POP · CAPEChamfer-L2 mean: 0.639 -> 0.592Improved
train one cross-outfit model on all outfits instead of per-outfit models (ablation)POP · CAPENormal diff. mean: 1.146 -> 1.115Improved
train one cross-outfit model on all outfits instead of per-outfit models (ablation)POP · ReSynthChamfer-L2 mean: 1.356 -> 1.366No clear effect
replace patch-based SCALE with dense continuous-feature point cloud (per-outfit)SCALE · CAPEChamfer-L2 mean: 0.721 -> 0.639Improved
replace patch-based SCALE with dense continuous-feature point cloud (per-outfit)SCALE · ReSynthChamfer-L2 mean: 1.491 -> 1.356Improved
replace implicit NASA with POP (per-outfit)NASA · CAPEChamfer-L2 mean: 6.087 -> 0.639Improved
train with half of the data per outfit (ablation)POP cross-outfit · CAPEChamfer-L2 mean: 0.592 -> 0.598Hurt
train with a quarter of the data per outfit (ablation)POP cross-outfit · CAPEChamfer-L2 mean: 0.592 -> 0.621Hurt
train with one eighth of the data per outfit (ablation)POP cross-outfit · CAPEChamfer-L2 mean: 0.592 -> 0.662Hurt
train with one eighth of the data per outfit (ablation)POP cross-outfit · ReSynthChamfer-L2 mean: 1.366 -> 1.490Hurt
use POP rather than SCALE judged by human ratersSCALE · CAPE visual qualityuser preference (%): 89.8% vs 10.2%Improved
use POP rather than SCALE judged by human ratersSCALE · ReSynth visual qualityuser preference (%): 84.8% vs 15.2%Improved

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