AI Paper Atlas

Research map / Video / Action recognition

S2Gest: Split-Scan State Space Models for Dynamic Hand Gesture Recognition

KeFan Chen, Yong Gu, bo li, Longjie Huang, Jiajun Zhang

ECCV 2026 · Video › Action recognition · approach: Skeleton-based action recognition · 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
use MS-Block spatial stem instead of 3D ResNet-18 (ablation)S2Gest-Small · NVGesture DepthTop-1 accuracy (%): 83.82 -> 90.66Improved
use S2-Block (Mamba) instead of Transformer encoder (ablation)S2Gest-Small · NVGesture DepthTop-1 accuracy (%): 86.51 -> 90.66Improved
add AFM guidance instead of no guidance (ablation)S2Gest-Small · NVGesture DepthTop-1 accuracy (%): 88.80 -> 90.66Improved
add naive element-wise addition of context instead of no guidance (ablation)S2Gest-Small · NVGesture DepthTop-1 accuracy (%): 88.80 -> 87.76Hurt
use AFM instead of concatenation of context (ablation)S2Gest-Small · NVGesture DepthTop-1 accuracy (%): 89.21 -> 90.66Improved
use Split-Scan instead of uni-directional scan (ablation)S2Gest-Small · NVGesture DepthTop-1 accuracy (%): 89.42 -> 90.66Improved
use Split-Scan instead of full-channel bi-directional scan (ablation)S2Gest-Small · NVGesture DepthTop-1 accuracy (%): 90.04 -> 90.66Improved
use Split-Scan instead of full bi-directional scan (parameters) (ablation)S2Gest-Small · NVGesture DepthParams (M): 3.26 -> 2.59Improved
use S2Gest-Nano instead of MobileNetV2lightweight backbone · NVGesture DepthTop-1 accuracy (%): 75.73 -> 82.99Improved
use depth modality instead of RGBS2Gest-Small · NVGestureTop-1 accuracy (%): 86.10 -> 90.66Improved
shrink model from Small to Nano (smaller capacity)S2Gest · Briareo RGBTop-1 accuracy (%): 97.69 -> 98.15Improved
shrink model from Small to NanoS2Gest · NVGesture DepthTop-1 accuracy (%): 90.66 -> 82.99Hurt

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