AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

LEO-Fuse: A Modality- and Task-Agnostic Universal Framework for Multimodal Human Sensing

Emrecan Aslan

ECCV 2026 · Vision-language and multimodal › Cross-modal learning · approach: Multi-modality image fusion

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 local routing head (ablation)LEO-Fuse · MM-Fi (31-subset average)MPJPE (mm): 88.31 -> 75.53Improved
add unimodal deep supervision (ablation)LEO-Fuse · MM-Fi (31-subset average)MPJPE (mm): 83.44 -> 75.53Improved
use K-based weighted subset sampling instead of uniform (ablation)LEO-Fuse · MM-Fi (31-subset average)MPJPE (mm): 79.54 -> 75.53Improved
add mode-aware loss balancing (ablation)LEO-Fuse · MM-Fi (31-subset average)MPJPE (mm): 78.75 -> 75.53Improved
add gate supervision (ablation)LEO-Fuse · MM-Fi (31-subset average)MPJPE (mm): 76.86 -> 75.53Improved
shrink pose head from 14.46M to 4.49M parameters (ablation)LEO-Fuse · MM-Fi (31-subset average)MPJPE (mm): 75.53 -> 76.02No clear effect
train one universal model across subsets vs dedicated LiDAR modelLEO-Fuse · MM-Fi LiDAR singletonMPJPE (mm): 121.02 -> 77.93Improved
train one universal model across subsets vs dedicated mmWave modelLEO-Fuse · MM-Fi mmWave singletonMPJPE (mm): 135.51 -> 119.97Improved
train one universal model vs dedicated WiFi modelLEO-Fuse · MM-Fi WiFi singletonMPJPE (mm): 199.09 -> 203.38Hurt
train frozen encoders with HAR labels instead of reusing HPE encodersLEO-Fuse · MM-Fi HAR (31-subset average)top-1 accuracy (%): 93.4 -> 95.7Improved
add more modalities (K=1 to K=4)LEO-Fuse · MM-Fi HPEMPJPE (mm): 106.21 -> 61.08Improved
add WiFi to D+L+R+I (full five modalities)LEO-Fuse · MM-Fi HPEMPJPE (mm): 57.22 -> 58.82Hurt

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