Poly-YOLO: higher speed, more precise detection and instance segmentation for YOLOv3

Poly-YOLO：改进YOLOv3实现速度更快、更准确的检测和分割.

Dynamic R-CNN: Towards High Quality Object Detection via Dynamic Training

Dynamic R-CNN：通过动态训练实现高质量目标检测.

Sparse R-CNN: End-to-End Object Detection with Learnable Proposals

Sparse R-CNN：基于可学习提议的端到端目标检测.

RepPoints: Point Set Representation for Object Detection

RepPoints：目标检测中的点集表示.

AutoAssign: Differentiable Label Assignment for Dense Object Detection

AutoAssign：密集目标检测中的可微标签分配.

Generalized Focal Loss V2: Learning Reliable Localization Quality Estimation for Dense Object Detection

Generalized Focal Loss V2：学习密集目标检测中可靠的定位质量估计.