CAD-Net: A Context-Aware Detection Network for Objects in Remote Sensing Imagery

By Gongjie Zhang, Shijian Lu, Wei Zhang. IEEE Transactions on Geoscience and Remote Sensing, 2019

424 citations recorded by the source corpus.

Accurate and robust detection of multi-class objects in optical remote sensing images is essential to many real-world applications, such as urban planning, traffic control, searching, and rescuing. However, the state-of-the-art object detection techniques designed for images captured using ground-level sensors usually experience a sharp performance drop when directly applied to remote sensing images, largely due to the object appearance differences in remote sensing images in terms of sparse texture, low contrast, arbitrary orientations, and large-scale variations. This paper presents a novel object detection network [(context-aware detection network (CAD-Net)] that exploits attention-modulated features as well as global and local contexts to address the new challenges in detecting objects from remote sensing images. The proposed CAD-Net learns global and local contexts of objects by capturing their correlations with the global scene (at scene level) and the local neighboring objects or features (at object level), respectively. In addition, it designs a spatial-and-scale-aware attention module that guides the network to focus on more informative regions and features as well as more appropriate feature scales. Experiments over two publicly available object detection data sets for remote sensing images demonstrate that the proposed CAD-Net achieves superior detection performance. The implementation codes will be made publicly available for facilitating future works.

Subfields: Computer Vision and Pattern Recognition, Media Technology

View Paper: https://www.semanticscholar.org/paper/69034a8f0329b6cb749e02ccee5ce199e9e06096

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

1. Junchi Yan
2. Xue Yang
3. Junwei Han
4. Gong Cheng
5. Jianke Zhu
6. Wentong Li
7. Guisong Xia
8. Jian Ding

In Media Technology

1. Gong Cheng
2. Junwei Han
3. Guisong Xia
4. Xiwen Yao
5. Nassim Ammour
6. Jian Ding
7. Xue Yang
8. Junchi Yan

Citations in the corpus

• Towards Large-Scale Small Object Detection: Survey and Benchmarks (2022, background): “Zhang et al. [137] captured the correlations between small objects and global scene (global context), as well as their neighboring instances (local context) to improve the performance.”
• SCRDet++: Detecting Small, Cluttered and Rotated Objects via Instance-Level Feature Denoising and Rotation Loss Smoothing (2020, methodology): “It is worth noting that FADet [27], SCRDet [30] and CAD-Det [25] use the simple attention mechanism by (1), but our performance is far better than all.”
• Classification of Remote Sensing Images Using EfficientNet-B3 CNN Model With Attention (2021, methodology): “In remote sensing, some works that use attention are proposed for RS object detection [37], RS image segmentation [38], [39], and RS scene classification [40]–[49].”
• R3Det: Refined Single-Stage Detector with Feature Refinement for Rotating Object (2019, result): “In contrast to the recent learning based methods (Chen et al. 2018; Jang et al. 2019; Zhang et al. 2019) for feature alignment, which lacks an explicit mechanism to compensate the misalignment, we propose a direct and effective pure computing based approach which is further extended to handle the ro”
• Dynamic Anchor Learning for Arbitrary-Oriented Object Detection (2020, background): “Some researches (Zhang et al. 2019; Song, Liu, and Wang 2020) have reported that the discriminative features required to localize objects are not evenly distributed on GT, especially for objects with a wide variety of orientations and aspect ratios.”
• ReDet: A Rotation-equivariant Detector for Aerial Object Detection (2021, methodology): “With the development of general object detection, many well-designed methods have been proposed for oriented object detection [1, 7, 24, 35, 38, 40, 42], showing promising performance on challenging datasets [21, 35].”
• Align Deep Features for Oriented Object Detection (2020, methodology): “3062048 To achieve better detection performance, most state-ofthe-art aerial object detectors [4], [5], [7], [8] rely on the complex R-CNN [9] frameworks, which consist of two parts: a region proposal network (RPN) and an R-CNN detection head.”
• FSDR: Frequency Space Domain Randomization for Domain Generalization (2021, background): “Deep learning based methods [4, 38, 79, 84] have achieved great successes at the price of large-scale densely-annotated training data [8] that are usually prohibitively expensive and timeconsuming to collect.”
• Dense Label Encoding for Boundary Discontinuity Free Rotation Detection (2020, methodology): “In the aerial imagery scene, ICN [2], ROITransformer [6], CAD-Net[54], SCRDet [51], R(3)Det [45], and CSL [48] achieve promising performance.”
• Oriented RepPoints for Aerial Object Detection (2021, background): “SCRDet [45], CADNet [47], DRN [24], R3Det [42], ReDet [9] and Oriented RCNN [35] achieve the promising performance by predicting the rotation angles of bounding boxes.”
• Gliding Vertex on the Horizontal Bounding Box for Multi-Oriented Object Detection (2019, background): “[28] propose to learn global and local contexts together to enhance the features.”