AI Paper Atlas

Research map / Video / Video generation

SFM: Taming State Space Models for Text-to-Motion via Spatial-Frequency Modeling

Shang Gao, Haicheng Liao, Wenshuo Chen, Yumu Xie, Jiaxun Zhang, Bin Rao, Chengyue Wang, Yanchen Guan, Zhiyong Cui, Shiqi Ou, Yutao Yue, Zhenning Li

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 frequency state-space module (F-SSM) (ablation)SFM · HumanML3D ablationR-Precision Top-3: 0.733 -> 0.845Improved
add spatial state-space module (S-SSM) (ablation)SFM · HumanML3D ablationR-Precision Top-3: 0.779 -> 0.845Improved
add semantic attention after dual-axis fusion (ablation)SFM · HumanML3D ablationR-Precision Top-3: 0.758 -> 0.845Improved
add text-gated frequency modulation (ablation)SFM · HumanML3D ablationR-Precision Top-3: 0.804 -> 0.845Improved
use soft text-gated adaptation instead of hard DCT coefficient selection (k=64) (ablation)SFM · HumanML3D ablationR-Precision Top-3: 0.587 -> 0.845Improved
use DCT instead of DWT, FFT or STFT (ablation)SFM · HumanML3D ablationR-Precision Top-3: FFT 0.815, DWT 0.834 -> DCT 0.845Improved
replace vanilla Mamba (Motion Mamba) with SFM dual-axis designSFM · HumanML3DFID: 0.281 -> 0.066Improved
SFM vs Transformer MoMask on FIDSFM · HumanML3DFID: 0.045 -> 0.066Hurt
SFM vs Transformer MDM on inference speedSFM · HumanML3D inference (A100)time per sentence: 15.33 s -> 0.19 sImproved
SFM vs CNN StableMoFusion on inference speedSFM · HumanML3D inference (A100)time per sentence: 0.13 s -> 0.19 sHurt
SFM vs Motion Mamba on long sequencesSFM · HumanML3D-LS (>190 frames)R-Precision Top-3: 0.713 -> 0.801Improved
SFM vs MoMask on long sequences FIDSFM · HumanML3D-LS (>190 frames)FID: 0.125 -> 0.190Hurt

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