Kim, Minsu Joung, Sunghun Song, Taeyong Kim, Hanjae Sohn, Kwanghoon 2024-01-19T10:03:38Z 2024-01-19T10:03:38Z 2023-03-23 2023-02 1545-598X https://pubs.kist.re.kr/handle/201004/114007 Object detection in aerial images has achieved remarkable progress with the advent of deep convolutional neural networks (CNNs). It is, however, still a challenging task since the objects in aerial images are arbitrarily oriented and often densely packed. In this letter, we propose a novel method for oriented object detection in aerial images that represents objects as rotation equivariant semantic keypoints. Unlike conventional methods that represent object rotation according to angles from each axis in the Cartesian coordinate system, we represent object using a canonical orientation to ensure rotation equivariance. We accomplish this by representing an object as semantic keypoints, where each keypoint of the object consistently corresponds to the semantic part, regardless of rotation variation. To this end, we define the "head " point of the object as the canonical orientation and the remaining bounding box vectors as semantic keypoints in clockwise order. To discriminate visual attributes between different categories, we further use category-specific semantic keypoints, so that object classification and localization can be jointly solved in a cooperative manner. Our experiments demonstrate the effectiveness of rotation equivariant semantic keypoints on oriented object detection. English Institute of Electrical and Electronics Engineers Learning Semantic Keypoints for Object Detection in Aerial Images Article 10.1109/LGRS.2022.3226201 1 IEEE Geoscience and Remote Sensing Letters, v.20 IEEE Geoscience and Remote Sensing Letters 20 N scie scopus 000966738000001 2-s2.0-85144074298 Geochemistry & Geophysics Engineering, Electrical & Electronic Remote Sensing Imaging Science & Photographic Technology Geochemistry & Geophysics Engineering Remote Sensing Imaging Science & Photographic Technology Article Semantics Object detection Feature extraction Location awareness Heating systems Head Image color analysis Convolutional neural networks (CNNs) equivariant representation oriented object detection remote sensing