AI Paper Atlas

Research map / Detection and segmentation / CNN features

Knowledge-Guided Adversarial Training for Infrared Object Detection via Thermal Radiation Modeling

Shiji Zhao, Shukun Xiong, Maoxun Yuan, Yao Huang, Ranjie Duan, Qing Guo, Jiansheng Chen, Haibin Duan, Xingxing Wei

IJCV 2026 · Detection and segmentation › CNN features · approach: Region-based and query-based detectors · 1 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 KGAT to MTD adversarial trainingMTD on Faster R-CNN · M3FDmAP50 (Avg): 49.9 -> 60.5Improved
add KGAT to SALC adversarial trainingSALC on DINO · M3FDmAP50 (Avg): 63.2 -> 70.8Improved
add KGAT to SALC adversarial trainingSALC on YOLO-v8 · M3FDmAP50 (Avg): 58.2 -> 66.5Improved
add KGAT to RobustDetRobustDet on SSD · M3FDmAP50 (Avg): 46.8 -> 50.0Improved
add KGAT to MTDMTD on SSD · M3FDmAP50 (Avg): 37.3 -> 43.2; clean 50.3 -> 56.1Improved
add KGAT to RobustDetRobustDet on SSD · M3FDmAP50 under A_mtd attack: A_mtd 37.9 -> 37.6Mixed
add KGAT to SALCSALC on Faster R-CNN · FLIR-ADASmAP50 (Avg): 49.6 -> 54.7Improved
add KGAT to SALCSALC on YOLO-v8 · FLIR-ADASmAP50 (Avg): 47.1 -> 53.4Improved
add KGAT to RobustDetRobustDet on SSD · FLIR-ADASmAP50 (Avg): 43.5 -> 45.6Improved
add KGAT to SALCSALC on Faster R-CNN · FLIR-ADASmAP50 (clean): 64.2 vs 62.3 clean (STD vs SALC+KGAT)Hurt
adversarial training instead of standard trainingSTD on DINO · M3FDmAP50 (clean): 86.9 -> 81.6 cleanHurt

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