AI Paper Atlas

Research map / Detection and segmentation / CNN features

SimROD: A Simple Baseline for Raw Object Detection with Global and Local Enhancements

Haiyang Xie, Xi Shen, Shihua Huang, Qirui Wang, Zheng Wang

AAAI 2026 · Detection and segmentation › CNN features · approach: Box regression and loss design · 2 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 green-channel guidance instead of no guidance in GGLE (ablation)SimROD (YoloX-Tiny) · LODAP50: 45.1 -> 46.3Improved
use red guidance in GGLE (ablation)SimROD · LODAP50: 45.1 -> 44.9Hurt
use blue guidance in GGLE (ablation)SimROD · LODAP50: 45.1 -> 44.2Hurt
guide with all RGGB channels instead of GG only (ablation)SimROD · LODAP50: 46.3 -> 46.2Hurt
guide with GG only vs RGGB on Pascal-Raw (ablation)SimROD · Pascal-RawAP50: 94.5 -> 95.1Improved
halve green channel sampling (0.5x) for DIAP (ablation)DIAP · LODAP50: 42.0 -> 40.3Hurt
halve green channel sampling on Pascal-Raw (ablation)DIAP · Pascal-RawAP50: 94.2 -> 94.3No clear effect
replace DIAP image-level module by GGE (4 params) (ablation)YoloX-Tiny · RODAP: 30.7 -> 30.9No clear effect
replace DIAP module by GGE on Pascal-Raw (ablation)YoloX-Tiny · Pascal-RawAP: 68.7 -> 68.4No clear effect
replace DIAP module by GGE on LOD (ablation)YoloX-Tiny · LODAP50: 43.4 -> 46.2Improved
SimROD vs DIAP on ROD with YoloX-TinySimROD · RODAP: 30.7 -> 33.1Improved
SimROD vs RAW-Adapter on over-exposed ADE20K-RawSimROD · ADE20K-Raw (over-exposure)mIoU: 46.62 -> 45.96Hurt

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