论文标题
PP-YOLOE-R：有效的无锚旋转对象检测器
PP-YOLOE-R: An Efficient Anchor-Free Rotated Object Detector
论文作者
论文摘要
在涉及空中图像和场景文本的视觉场景中，任意面向对象检测是一项基本任务。在本报告中，我们提出了PP-Yoloe-R，这是一种基于PP-Yoloe的有效无锚旋转的对象检测器。我们在PP-Yoloe-R中引入了一袋有用的技巧，以通过边际额外参数和计算成本提高检测精度。结果，PP-Yoloe-R-L和PP-Yoloe-R-X分别在DOTA 1.0数据集上分别实现了78.14和78.28的地图，并具有单尺度训练和测试，几乎所有其他旋转的对象探测器的表现几乎超过了所有其他旋转的对象探测器。通过多尺度训练和测试，PP-Yoloe-R-L和PP-Yoloe-R-X进一步提高了检测精度为80.02和80.73地图。在这种情况下，PP-Yoloe-R-X超过了所有无锚方法，并向基于最新的锚固型两阶段模型展示了竞争性能。此外，PP-Yoloe-R在RTX 2080 TI上分别达到69.8/55.1/48.3/37.1 fps的PP-Yoloe-R-S/M/L/X可以分别达到69.8/55.1/48.3/37.1 fps。源代码和预训练模型可在https://github.com/paddlepaddle/paddledetection上获得，该模型由https://github.com/paddledlepaddle/paddle提供。
Arbitrary-oriented object detection is a fundamental task in visual scenes involving aerial images and scene text. In this report, we present PP-YOLOE-R, an efficient anchor-free rotated object detector based on PP-YOLOE. We introduce a bag of useful tricks in PP-YOLOE-R to improve detection precision with marginal extra parameters and computational cost. As a result, PP-YOLOE-R-l and PP-YOLOE-R-x achieve 78.14 and 78.28 mAP respectively on DOTA 1.0 dataset with single-scale training and testing, which outperform almost all other rotated object detectors. With multi-scale training and testing, PP-YOLOE-R-l and PP-YOLOE-R-x further improve the detection precision to 80.02 and 80.73 mAP. In this case, PP-YOLOE-R-x surpasses all anchor-free methods and demonstrates competitive performance to state-of-the-art anchor-based two-stage models. Further, PP-YOLOE-R is deployment friendly and PP-YOLOE-R-s/m/l/x can reach 69.8/55.1/48.3/37.1 FPS respectively on RTX 2080 Ti with TensorRT and FP16-precision. Source code and pre-trained models are available at https://github.com/PaddlePaddle/PaddleDetection, which is powered by https://github.com/PaddlePaddle/Paddle.
