English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Object detection on visible (RGB) and infrared (IR) images, as an emergingsolution to facilitate robust detection for around-the-clock applications, hasreceived extensive attention in recent years. With the help of IR images,object detectors have been more reliable and robust in practical applicationsby using RGB-IR combined information. However, existing methods still sufferfrom modality miscalibration and fusion imprecision problems. Since transformerhas the powerful capability to model the pairwise correlations betweendifferent features, in this paper, we propose a novel Calibrated andComplementary Transformer called $\mathrm{C}^2$Former to address these twoproblems simultaneously. In $\mathrm{C}^2$Former, we design an Inter-modalityCross-Attention (ICA) module to obtain the calibrated and complementaryfeatures by learning the cross-attention relationship between the RGB and IRmodality. To reduce the computational cost caused by computing the globalattention in ICA, an Adaptive Feature Sampling (AFS) module is introduced todecrease the dimension of feature maps. Because $\mathrm{C}^2$Former performsin the feature domain, it can be embedded into existed RGB-IR object detectorsvia the backbone network. Thus, one single-stage and one two-stage objectdetector both incorporating our $\mathrm{C}^2$Former are constructed toevaluate its effectiveness and versatility. With extensive experiments on theDroneVehicle and KAIST RGB-IR datasets, we verify that our method can fullyutilize the RGB-IR complementary information and achieve robust detectionresults. The code is available athttps://github.com/yuanmaoxun/Calibrated-and-Complementary-Transformer-for-RGB-Infrared-Object-Detection.git.

$\mathbf{C}^2$Former: Calibrated and Complementary Transformer for RGB-Infrared Object Detection