Abstract
In remote sensing images, the backgrounds of objects include crucial contextual information that may contribute to distinguishing objects. However, there are at least two issues that should be addressed: not all the backgrounds are beneficial, and object information may be suppressed by backgrounds. To address these problems, in this article, we propose the contextual bidirectional enhancement (CBD-E) method to simultaneously remove unexpected background information and enhance objects’ features. CBD-E integrates the features of different background regions sequentially in two directions. On the one hand, a gate function is used to filter out unexpected information in the background and thus improve the recall of detection. On the other hand, a spatial-group-based visual attention mechanism is adopted to enhance the features of objects to reduce the false alarm. The gate function provides an approach to selecting meaningful information in the background, while the spatial-group- based visual attention mechanism enhances the information control ability of the gate function. In the experiments, we have validated the effectiveness of both the gate function and the visual attention mechanism and further demonstrated that the proposed contextual fusion strategy performs well on two published data sets.
Highlights
O BJECT detection in remote sensing images has attracted more and more attention and has achieved remarkable results in recent years [1]–[4]
The fused feature maps are fed into the parameter sharing detection network to generate multiple prediction boxes, and the optimal box is selected by nonmaximum suppression (NMS)
Instead of retaining and passing all the information of the background to the region during the fusion process, we introduce gate functions to filter out unexpected information and adapt message passing for individual candidate boxes
Summary
O BJECT detection in remote sensing images has attracted more and more attention and has achieved remarkable results in recent years [1]–[4]. The objects and the background in remote sensing images have certain contextual relationship. We construct a contextual-based remote sensing object detection network called contextual bidirectional enhancement (CBD-E). CBD-E is motivated by the idea that context generally contribute to the detection results, but under certain circumstances the background may mislead the detectors. We integrate the gated bidirectional fusion (GBD) structure [10] into our CBD-E model to suppress the unexpected background. To further enhance the objects in the image, we improve the bidirectional fusion structure via a visual attention based approach, spatial groupwise enhancement (SGE) [11]. ZHANG et al.: CONTEXTUAL BIDIRECTIONAL ENHANCEMENT METHOD FOR REMOTE SENSING IMAGE OBJECT DETECTION. 2) We conduct SGE to improve the object saliency and highlight the features of the object
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
