AI Paper Atlas

Research map / Detection and segmentation / CNN features

TinyViM: Frequency Decoupling for Tiny Hybrid Vision Mamba

Xiaowen Ma, Zhenliang Ni, Xinghao Chen

ICCV 2025 · Detection and segmentation › CNN features · approach: Vision state space models · 42 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 Mamba global modeling to a conv-only hybrid (ablation)TinyViM-S layout · ImageNet-1K val (Mamba input ablation, TinyViM-S layout)top-1 accuracy (%): 77.5 -> 79.1Improved
feed only low-frequency components to Mamba instead of full features (ablation)TinyViM-S layout · ImageNet-1K val (Mamba input ablation, TinyViM-S layout)top-1 accuracy (%): 79.1 -> 79.0No clear effect
feed only low-frequency components to Mamba instead of full features (ablation)TinyViM-S layout · ImageNet-1K val (Mamba input ablation, TinyViM-S layout)throughput (im/s): 1673 -> 2574Improved
feed only high-frequency input to Mamba instead of low only (ablation)TinyViM-S layout · ImageNet-1K val (Mamba input ablation, TinyViM-S layout)top-1 accuracy (%): 79.0 -> 78.6Hurt
add high-frequency input in parallel to Mamba on top of low (ablation)TinyViM-S layout · ImageNet-1K val (Mamba input ablation, TinyViM-S layout)throughput (im/s): 2574 -> 1509Hurt
add frequency ramp (alpha 0.25/0.5/0.5/0.75) to inception (ablation)TinyViM-S · ImageNet-1K valtop-1 accuracy (%): 79.0 -> 79.2Improved
add frequency ramp inception versus decoupling all channels (ablation)TinyViM-S · ImageNet-1K valthroughput (im/s): 2478 -> 2563Improved
use ramp alpha increasing with depth instead of inverse ramp (ablation)TinyViM-S · ImageNet-1K valtop-1 accuracy (%): 78.8 -> 79.2Improved
use axial depthwise kernel 7 instead of 3x3 depthwise (ablation)TinyViM-S · ImageNet-1K valtop-1 accuracy (%): 79.0 -> 79.2Improved
increase axial kernel from 7 to 9 (ablation)TinyViM-S · ImageNet-1K valtop-1 accuracy (%): 79.2 -> 79.2No clear effect
add re-parameterization (ablation)TinyViM-L · ImageNet-1K valtop-1 accuracy (%): 83.1 -> 83.3Improved
add re-parameterization (ablation)TinyViM-S · ImageNet-1K valtop-1 accuracy (%): 79.2 -> 79.2No clear effect

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