AI Paper Atlas

Research map / Detection and segmentation / 3D detection and driving

Beyond Appearance: Camouflaged Object Detection via Geometric Structure

Jinyu Han, Changguang Wu, Fuming Sun, Jinhui Tang

CVPR 2026 · Detection and segmentation › 3D detection and driving · approach: Monocular 3D object detection

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 Sparse Mixture-of-Experts Adapter (SMEA) to frozen MDE encoder (ablation)DepthSAM-L · COD10Kweighted F-measure: 0.719 -> 0.867Improved
use sparse multi-expert SMEA instead of a standard adapter (ablation)DepthSAM-L · COD10Kweighted F-measure: 0.840 -> 0.867Improved
increase activated experts from Top-K=2 to Top-K=4 (ablation)DepthSAM-L · COD10Kweighted F-measure: 0.867 -> 0.869No clear effect
add Geometric-Semantic Fusion Module (GSFM) instead of simple concatenation (ablation)DepthSAM-L · COD10Kweighted F-measure: 0.766 -> 0.867Improved
use GSFM instead of a standard FFN for fusion (ablation)DepthSAM-L · COD10Kweighted F-measure: 0.829 -> 0.867Improved
use spatial-frequency SFRM instead of standard MHSA in GSFM (ablation)DepthSAM-L · COD10Kweighted F-measure: 0.809 -> 0.867Improved
use Depth Anything v2 backbone instead of SAM with the same components (ablation)DepthSAM-L · COD10KS-measure: 0.880 -> 0.920Improved
increase input resolution from 384 to 512 (ablation)DepthSAM-L · COD10KS-measure: 0.905 -> 0.920Improved
apply COD-trained model zero-shot to video camouflaged detectionDepthSAM-L · MoCA-MaskS-measure: 73.4 -> 78.4 vs ZoomNeXtImproved
retrain DepthSAM-L on RGB salient object detectionDepthSAM-L · DUTSS-measure: 0.932 -> 0.947 vs SambaImproved
retrain DepthSAM-L on optical remote sensing SODDepthSAM-L · EORSSDMAE: S 0.9449 -> 0.9503; MAE 0.0049 (HFCNet) vs 0.0053Mixed

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