.png){kind=icon}
Return
IRDFusion: Iterative relation-map difference guided feature fusion for multispectral object detection
.png){kind=icon}
PRE
AI
DOI：10.1016/j.patcog.2026.113189.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• A Mutual Feature Refinement Module (MFRM) is proposed to enhance modalspecific features of object candidates between two modalities, ensuring robust feature alignment. • Inspired by the feedback differential amplifier circuits, a Differential Feature Feedback Module (DFFM) is proposed to calculate complementary discriminative features between the two modalities and simultaneously filters redundant information. • The MFRM and DFFM are jointly optimized to effectively integrate discriminative complementary information from different modalities through a dynamic differential relationship map feedback mechanism, which provides a new strategy for progressive multispectral feature fusion. • The proposed method IRDFusion, building on MFRM and DFFM, achieves state-of-the-art performance on the FLIR, LLVIP and M3FD dataset.
Journal
IF:
7.6
Papers:
1.3W
Citations:
4.5W
Organization
Cited Papers
PIAFusion: A progressive infrared and visible image fusion network based on illumination aware.png)
INFORMATION FUSION
INFORMATION FUSIONIF15.5
PREAI
ShareSaveImproving RGB-infrared object detection with cascade alignment-guided transformerINFORMATION FUSION
INFORMATION FUSIONIF15.5
PREAI
ShareSaveShareSaveShareSaveIllumination-aware faster R-CNN for robust multispectral pedestrian detectionPATTERN RECOGNITION
OAAI
PATTERN RECOGNITIONIF7.6
OAAIShareSaveShareSaveShareSave
View more