R3Det: Refined Single-Stage Detector with Feature Refinement for Rotating Object
By Xue Yang, Qingqing Liu, Junchi Yan, Ang Li. AAAI Conference on Artificial Intelligence, 2019
1,161 citations recorded by the source corpus.
Rotation detection is a challenging task due to the difficulties of locating the multi-angle objects and separating them effectively from the background. Though considerable progress has been made, for practical settings, there still exist challenges for rotating objects with large aspect ratio, dense distribution and category extremely imbalance. In this paper, we propose an end-to-end refined single-stage rotation detector for fast and accurate object detection by using a progressive regression approach from coarse to fine granularity. Considering the shortcoming of feature misalignment in existing refined single-stage detector, we design a feature refinement module to improve detection performance by getting more accurate features. The key idea of feature refinement module is to re-encode the position information of the current refined bounding box to the corresponding feature points through pixel-wise feature interpolation to realize feature reconstruction and alignment. For more accurate rotation estimation, an approximate SkewIoU loss is proposed to solve the problem that the calculation of SkewIoU is not derivable. Experiments on three popular remote sensing public datasets DOTA, HRSC2016, UCAS-AOD as well as one scene text dataset ICDAR2015 show the effectiveness of our approach. The source code is available at https://github.com/Thinklab-SJTU/R3Det_Tensorflow and is also integrated in our open source rotation detection benchmark: https://github.com/yangxue0827/RotationDetection.
Subfields: Aerospace Engineering, Archeology, Computer Vision and Pattern Recognition
View Paper: https://www.semanticscholar.org/paper/edada2363969e3929366df06aad8a8e9c73ba32f
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Aerospace Engineering
In Archeology
In Computer Vision and Pattern Recognition
Citations in the corpus
- Oriented RepPoints for Aerial Object Detection (2021, background): “SCRDet [45], CAD-Net [47], DRN [24], R3Det [42], ReDet [9] and Oriented R-CNN [36] achieve the promising performance by predicting the rotation angles of bounding boxes.”
- Align Deep Features for Oriented Object Detection (2020, background): “Recently proposed R3Det [24] samples features from five locations (e.g., center and corners) of the corresponding AB and sum them up to re-encode the position information.”
- SCRDet++: Detecting Small, Cluttered and Rotated Objects via Instance-Level Feature Denoising and Rotation Loss Smoothing (2020, methodology): “04) ICDAR2015 [28] RetinaNet-H [18] 77.”
- RTMDet: An Empirical Study of Designing Real-Time Object Detectors (2022, background): “, RetinaNet [46] or Faster R-CNN [65]), different feature extraction networks are proposed to alleviate the feature misalignment [28, 29, 87] caused by object rotations.”
- Advancing Plain Vision Transformer Toward Remote Sensing Foundation Model (2022, neutral): “92 Multi-scale R(3)Det [47] IMP ResNet-152 89.”
- Oriented R-CNN for Object Detection (2021, background): “al [39] proposed a refined one-stage oriented object detector, which involves two key improvements, including feature refinement and progressive regression, to address the problem of feature misalignment.”
- 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.”
- Localization Distillation for Object Detection (2022, methodology): “Parametric regression is the most popular manner in the classical dense regression-based rotated object detection [71], [73], [74], [75].”
- Dynamic Anchor Learning for Arbitrary-Oriented Object Detection (2020, methodology): “R(3)Det(Yang et al. 2019a) adopts cascade regression and refined box re-encoding module to achieve state-of-the-art performance with horizontal”
- PIoU Loss: Towards Accurate Oriented Object Detection in Complex Environments (2020, methodology): “Existing OBB-based approaches are mostly built on anchor-based frameworks by introducing an additional angle dimension optimized by a distance loss [24,18,19,6,41,43] on the parameter tuple (cx, cy, w, h, θ).”
- Arbitrary-Oriented Object Detection with Circular Smooth Label (2020, background): “Three rotation detectors in Table 3, including RetinaNet-H, RetinaNet-R and FPN-H, are used to compare the performance differences between CSL-based and regression-based methods.”
- Dense Label Encoding for Boundary Discontinuity Free Rotation Detection (2020, methodology): “Many rotation detectors (including those for aerial image [46, 15, 47, 2, 6, 51, 45, 30, 48, 49, 50], scene text [56, 23, 14, 26, 19, 18] and face [34, 13, 33]) are derived based on the vanilla detection algorithms [10, 32, 31, 5, 20, 21].”
- MMRotate: A Rotated Object Detection Benchmark using PyTorch (2022, methodology): “R3Det [35] proposes a feature refinement module to re-construct the feature map based on the refined bounding box output from the previous stage.”
- ReDet: A Rotation-equivariant Detector for Aerial Object Detection (2021, background): “R3Det and S2A-Net align the feature between horizontal receptive fields and rotated anchors.”
- Center-based 3D Object Detection and Tracking (2020, background): “One solution might be to classify a different template (anchor) for each object orientation [56, 57], but this unnecessarily increases the computational burden and may introduce a large number of potential false-positive detections.”