地址:https://arxiv.org/abs/2006.04388 github:https://github.com/implus/GFocal 讲解:https://zhuanlan.zh ...
目录 Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection 一. 论文简介 将目标检测Loss ...
2020-10-23 11:50 0 413 推荐指数:
地址:https://arxiv.org/abs/2006.04388 github:https://github.com/implus/GFocal 讲解:https://zhuanlan.zh ...
Focal Loss for Dense Object Detection Intro 这又是一篇与何凯明大神有关的作品,文章主要解决了one-stage网络识别率普遍低于two-stage网络的问题,其指出其根本原因是样本类别不均衡导致,一针见血,通过改变传统的loss(CE)变为focal ...
转自:https://www.jianshu.com/p/db4ccd194109 ...
论文原址:https://arxiv.org/abs/1708.02002 github代码:https://github.com/fizyr/keras-retinanet 摘要 ...
论文地址:Generalized Intersection over Union 一、相关工作 目标检测精度标准 度量检测优劣基本基于 IOU,mAP 是典型的基于 IOU 的标准,但是 mAP 仅有一个 threshold,对于过了线的预测框一视同仁,不能进一步衡量其优劣,所以 MS ...
地址:https://arxiv.org/pdf/2007.03282.pdf ...
/Yemiekai/article/details/117839686 文章标题:《Weighted Boxes ...
论文原址:https://arxiv.org/pdf/1902.09630.pdf github:https://github.com/generalized-iou 摘要 在目标检测的评测体系中,IoU是最流行的评价准则。然而,在对边界框的参数进行优化时,常用到距离 ...