Align Deep Features for Oriented Object Detection
By Jiaming Han, Jian Ding, Jie Li, Guisong Xia. IEEE Transactions on Geoscience and Remote Sensing, 2020
1,122 citations recorded by the source corpus.
The past decade has witnessed significant progress on detecting objects in aerial images that are often distributed with large-scale variations and arbitrary orientations. However, most of existing methods rely on heuristically defined anchors with different scales, angles, and aspect ratios, and usually suffer from severe misalignment between anchor boxes (ABs) and axis-aligned convolutional features, which lead to the common inconsistency between the classification score and localization accuracy. To address this issue, we propose a single-shot alignment network (S2A-Net) consisting of two modules: a feature alignment module (FAM) and an oriented detection module (ODM). The FAM can generate high-quality anchors with an anchor refinement network and adaptively align the convolutional features according to the ABs with a novel alignment convolution. The ODM first adopts active rotating filters to encode the orientation information and then produces orientation-sensitive and orientation-invariant features to alleviate the inconsistency between classification score and localization accuracy. Besides, we further explore the approach to detect objects in large-size images, which leads to a better trade-off between speed and accuracy. Extensive experiments demonstrate that our method can achieve the state-of-the-art performance on two commonly used aerial objects’ data sets (i.e., DOTA and HRSC2016) while keeping high efficiency.
Subfields: Computer Vision and Pattern Recognition
View Paper: https://www.semanticscholar.org/paper/1360066463544a40c5dafd70152a85c437b4bf84
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
Citations in the corpus
- ReDet: A Rotation-equivariant Detector for Aerial Object Detection (2021, methodology): “R3Det and S2A-Net align the feature between horizontal receptive fields and rotated anchors.”
- Towards Large-Scale Small Object Detection: Survey and Benchmarks (2022, background): “S 2 A-Net [178] designs feature alignment module to alleviate the misalignment problem, and finally achieves an AP with 28.3%.”
- MMRotate: A Rotated Object Detection Benchmark using PyTorch (2022, neutral): “…OpenCV definition ( 𝐷 𝑜𝑐 ,𝜃 ∈ [− 𝜋 / 2 , 0 ) ) [32, 36, 38], long edge 90° definition ( 𝐷 𝑙𝑒 90 ,𝜃 ∈ [− 𝜋 / 2 , 𝜋 / 2 ) ) [3, 6], and long edge 135° ) ) [5]. https://github.com/dingjiansw101/AerialDetection 4 https://github.com/jbwang1997/OBBDetection Recent works used two-dimensional…”
- Dynamic Anchor Learning for Arbitrary-Oriented Object Detection (2020, background): “S(2)A-Net (Han et al. 2020) is an advanced rotation detector that achieves the state-of-the-art performance on DOTA dataset.”
- Poly Kernel Inception Network for Remote Sensing Detection (2024, methodology): “Top : Our approach yields solid performance gains over various remote sensing detectors [10, 20, 59, 65, 71] with fewer parameters on DOTA-v1.”
- Large Selective Kernel Network for Remote Sensing Object Detection (2023, methodology): “8, 9, 10), similarly to [6,23,69,75].”
- Object Detection in Aerial Images: A Large-Scale Benchmark and Challenges (2021, background): “The later S(2)A-Net [57] extracts spatially invariant features in onestage detectors.”
- Oriented R-CNN for Object Detection (2021, background): “As an extended branch of object detection, oriented object detection [7, 37, 26, 12, 5, 28] has received extensive attention driven by its wide applications.”
- Anchor-Free Oriented Proposal Generator for Object Detection (2021, background): “RRPN [6] R2CNN [41] RT [1] GV [2] DRN [21] PIoU [38] DAL [39] S(2)ANet [4] AOPG”
- Oriented RepPoints for Aerial Object Detection (2021, methodology): “The mainstreamed approaches typically treat aerial object detection as a problem of rotated object localization [3, 8,9,24,42,45].”
- Advancing Plain Vision Transformer Toward Remote Sensing Foundation Model (2022, result): “We search for the optimal value of s in the range of [4, 14] and the results are shown in Table III.”
- Learning High-Precision Bounding Box for Rotated Object Detection via Kullback-Leibler Divergence (2021, methodology): “The latter is based on the length and width of the anchor (wa, ha) to adjust the gradient instead of the target object (wt, ht), which is almost ineffective for those detectors [3, 13, 15, 27, 28, 32, 33] that use horizontal anchors for rotation detection.”