Sparse R-CNN: End-to-End Object Detection with Learnable Proposals
Sparse R-CNN: End-to-End Object Detection with Learnable Proposals目标檢測交流群 作者機關：港大, 同濟大學, 位元組AI Lab, UC伯克利
沿着目标檢測領域中Dense和Dense-to-Sparse的架構，Sparse R-CNN建立了一種徹底的Sparse架構， 脫離anchor box，reference point，Region Proposal Network(RPN)等概念，無需Non-Maximum Suppression(NMS)後處理， 在标準的COCO benchmark上使用ResNet-50 FPN單模型在标準3x training schedule達到了44.5 AP和 22 FPS。
- 解讀：https://zhuanlan.zhihu.com/p/310058362
- 代碼：https://github.com/PeizeSun/SparseR-CNN
- 論文：https://msc.berkeley.edu/research/autonomous-vehicle/sparse_rcnn.pdf
Sparse R-CNN: End-to-End Object Detection with Learnable Proposals目标檢測交流群 Sparse R-CNN: End-to-End Object Detection with Learnable Proposals目标檢測交流群 Sparse R-CNN: End-to-End Object Detection with Learnable Proposals目标檢測交流群 Sparse R-CNN: End-to-End Object Detection with Learnable Proposals目标檢測交流群 Sparse R-CNN: End-to-End Object Detection with Learnable Proposals目标檢測交流群 Sparse R-CNN: End-to-End Object Detection with Learnable Proposals目标檢測交流群 Sparse R-CNN: End-to-End Object Detection with Learnable Proposals目标檢測交流群 Sparse R-CNN: End-to-End Object Detection with Learnable Proposals目标檢測交流群 Sparse R-CNN: End-to-End Object Detection with Learnable Proposals目标檢測交流群 Sparse R-CNN: End-to-End Object Detection with Learnable Proposals目标檢測交流群 目标檢測交流群
CVer已建立目标檢測微信交流群，已經上千人！想要進目标檢測學習群的同學，可以直接加微信号：CVer5555。加的時候備注一下：目标檢測+學校+昵稱，即可。然後就可以拉你進群了。
Sparse R-CNN: End-to-End Object Detection with Learnable Proposals目标檢測交流群 
