AI Paper Atlas

Research map / Video / Video generation

K-Mask: Kinematic-Aware Masked Modeling for Controllable Text-to-Motion Synthesis

Faisal Ahmed, Chenqiu Zhao, Anup Basu

ECCV 2026 · Video › Video generation · approach: Text-to-motion diffusion · 0 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
add kinematic group masking (pkin 0.25) to random token masking (ablation)K-Mask Base-T · HumanML3DFID: 0.108 -> 0.041Improved
raise kinematic masking probability to 0.5 (ablation)K-Mask Base-T · HumanML3DFID: 0.041 -> 0.078Hurt
add LAKD loss to the group tokenizer (ablation)KG-RVQ · HumanML3D short-run tokenizercross-group leakage: 2.484 -> 0.088Improved
add LAKD loss to the group tokenizer (ablation)KG-RVQ · HumanML3D short-run tokenizerMPJPE: 16.4 -> 17.2Hurt
use 6 anatomical groups instead of whole-body tokenization (ablation)KG-RVQ · HumanML3D short-run tokenizerMPJPE: 30.6 -> 17.2Improved
use 22 joint-level groups instead of 6 (ablation)KG-RVQ · HumanML3D short-run tokenizercross-group leakage: 0.088 -> 1.489Hurt
proportional latent allocation instead of uniform split (ablation)KG-RVQ · HumanML3D short-run tokenizerMPJPE: 18.5 -> 17.2Improved
temporal-then-spatial factorized attention instead of spatial-then-temporal (ablation)K-Mask transformers · HumanML3DFID: 0.077 -> 0.041Improved
axis-factorized attention instead of full self-attention (ablation)K-Mask transformers · HumanML3DFID: 0.049 -> 0.041No clear effect
group-local token resampling for localized editing vs whole-body inpainting (MoMask)K-Mask · HumanML3D 200-case editingSpillover: .164 -> .108Improved
K-Mask vs MoGenTS on generation fidelityK-Mask · HumanML3D testFID: MoGenTS .028 vs K-Mask .041 (Table 2) while Table 1 text claims best FIDMixed

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 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