AI Paper Atlas

Research map / Detection and segmentation / CNN features

Efficient RGB-T Object Detection via Sparse Cross-Modality Fusion

CHAO TIAN, Zikun Zhou, Chao Yang, guoqing zhu, Zhenyu He

ECCV 2026 · Detection and segmentation › CNN features · approach: RGB-D salient object detection · 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 differential enhancement to group-wise fusion (ablation)FER fusion · M3FDmAP: 58.4 -> 58.6Improved
add group-wise fusion to differential enhancement (ablation)FER fusion · M3FDmAP: 58.1 -> 58.6Improved
add auxiliary backbone features (ablation)FER · M3FDAP50: 87.2 -> 88.5Improved
take auxiliary features from backbone instead of RPN features (ablation)FER · M3FDAP50: 88.1 -> 88.5Improved
add second refinement step (ablation)FER · M3FDAP50: 88.5 -> 89.8Improved
use rolling convolution instead of Conv2d (ablation)FER · M3FDAP50: 88.9 -> 89.8Improved
add denoising training (ablation)FER · M3FDAP50: 88.0 -> 89.8Improved
use rolling convolution instead of Shuffle (G=8) (ablation)FER · M3FDAP50: 88.9 -> 89.8Improved
use RoI size 5 instead of 3 (ablation)FER · M3FDAP50: 88.7 -> 89.8Improved
use RoI size 7 instead of 5 (ablation)FER · M3FDAP50: 89.8 -> 88.9Hurt
use 16 rolling groups instead of 8 (ablation)FER · M3FDAP50: 89.8 -> 88.9Hurt
sparse fusion instead of dense FPN fusion with DEGF (ablation)SFEDet · M3FDAP50: 88.4 -> 89.8Improved

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