AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

On Model and Data Scaling for Skeleton-based Self-Supervised Gait Recognition

Adrian Cosma, Andy Eduard Catruna, Emilian Radoi

AAAI 2026 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Invariance-based self-supervised learning · 3 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
select top-quality skeleton sequences by confidence-based quality score instead of random subsets (ablation)GaitPTv2 (mu=2) · controlled gait benchmarksdata scaling exponent beta: 0.078 -> 0.1319Improved
select top-quality skeleton sequences instead of random subsets (ablation)GaitPTv2 (mu=2) · in-the-wild gait benchmarksdata scaling exponent beta: 0.099 -> 0.125Improved
scale pretraining data to the full 2.7M sequencesGaitPTv2 · controlled gait benchmarksmodel scaling exponent alpha: 0.100 -> 0.164Improved
scale pretraining data from 43.4K to 2.78M sequencesGaitPTv2 · in-the-wild gait benchmarksmodel scaling exponent alpha: 0.118 -> 0.112No clear effect
increase model sizeGaitPTv2 · controlled and in-the-wild gait benchmarkszero-shot accuracyImproved
increase pretraining data to the largest scaleGaitPTv2 · gait benchmarkszero-shot accuracyMixed
use hierarchical GaitPTv2 instead of GaitFormer under same parameter countgait encoder · controlled gait recognitionaccuracyImproved
use hierarchical GaitPTv2 instead of GaitFormergait encoder · controlled gait recognitioncompute efficiency (FLOPs): 2 -> 6.5Hurt
increase parameter count at fixed compute (IsoFLOPs)GaitPTv2 and GaitFormer · controlled gait recognitionaccuracyHurt
scale parameter count of GaitPTv2 vs GaitFormer using compute-efficient modelsgait encoder · controlled gait recognitionaccuracyNo clear effect
scale number of training sequences for GaitPTv2 vs GaitFormergait encoder · controlled gait recognitionaccuracyImproved

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