目标检测
三寸光阴___
这个作者很懒,什么都没留下…
展开
-
基于Anchor Free的目标检测
总结:https://blog.csdn.net/weixin_36670529/article/details/1064688341.UnitBox: An Advanced Object Detection Network论文:代码:解析:2.DenseBox: Unifying Landmark Localization and Object Detection论文:代码:解析:3.You Only Look Once (YOLO) for Object Detection (YOL原创 2020-12-24 17:07:14 · 308 阅读 · 1 评论 -
域自适应目标检测
总结:https://www.yuque.com/weijiawu/research/vp3v2yAdversarial-based1.Unsupervised Domain Adaptation by Backpropagation论文:https://arxiv.org/abs/1409.7495代码:https://github.com/fungtion/DANN_py3解析:https://blog.csdn.net/weixin_37993251/article/details/8935原创 2020-12-14 17:08:03 · 3397 阅读 · 0 评论 -
YOLOv3模型剪枝
参考代码:https://github.com/Lam1360/YOLOv3-model-pruninghttps://github.com/coldlarry/YOLOv3-complete-pruning剪枝代码解读:https://blog.csdn.net/just_sort/article/details/107089710剪枝原理:https://openaccess.thecvf.com/content_iccv_2017/html/Liu_Learning_Efficient_Conv原创 2020-07-09 18:53:42 · 5718 阅读 · 1 评论 -
CenterNet(objects as points)算法总结:Mimic、DCNv1、DCNv2、DLA
DCNv1参考文章:https://zhuanlan.zhihu.com/p/37578271主要思路:卷积操作不是在规规矩矩的3x3的格子里做了,而是有了种种偏移。deformable convolution在特征图的每个位置都生成一组偏移量。以kernel=3x3为例,图中的N就是9,经过卷积后输出HxWx18的feature map,每个location位置的18分别对应3x3卷积...原创 2020-03-05 14:15:39 · 1802 阅读 · 0 评论 -
大论文资料收集(人体姿态估计)
自己总结:https://blog.csdn.net/qq_38109843/article/details/102396365总结:https://blog.csdn.net/qq_36165459/article/details/78320535?locationNum=10&fps=1http://www.sohu.com/a/345995744_99979179...原创 2019-10-29 21:17:19 · 187 阅读 · 0 评论 -
CornerNet-Lite论文笔记
论文链接:https://arxiv.org/pdf/1904.08900.pdf代码链接:https://github.com/princeton-vl/CornerNet-LiteCornerNet关于CornerNet论文解读在这里:https://blog.csdn.net/qq_38109843/article/details/89526785CornerNet主要思路是基于关键...原创 2019-06-05 10:55:36 · 1738 阅读 · 0 评论 -
tensorflow lite将ssd_mobilenet移植至移动端
准备工作下载tensorflow源码git clone https://github.com/tensorflow/tensorflow.git下载tensorflow modelgit clone https://github.com/tensorflow/models.git下载并安装bazel编译工具https://docs.bazel.build/versions/mast...原创 2019-08-09 10:27:27 · 1418 阅读 · 3 评论