1999:SIFT
2001:Cascades(LBP与HAAR特征)
2003:Bag of Words
2005:HOG
2006:SPM/SURF/Region Covariance
2007:PASCAL VOC
2008:DPM/Efficient Subwindow Search
2009:HOG-LBP/ImageNet
2010:Improved FV
2011:Selective Search
--------深度学习兴起,机器学习时代来临----------
2012:DCNN AlexNet(Alex和他的导师Hinton:83.6%的Top5精度)
2013:OverFeat (ImageNet:88.8%)
2014:MS COCO/RCNN (VGG的92.7%和同年的GoogLeNet的93.3%)
2015:Fast RCNN/Faster RCNN/VGGNet/GoogLeNet(微软的ResNet:96.43%的Top5正确率):人类的正确率也只有94.9%
2016:ResNet/SSD
2017:Mask RCNN/DenseNet
---------ImageNet大赛关闭-------------------
-------- 其他----------------------------------
2018CVPR目标检测论文:https://blog.csdn.net/wfei101/article/details/80861681
2019CVPR目标检测论文:https://blog.csdn.net/xiao_lxl/article/details/95621146
https://blog.csdn.net/f290131665/article/details/81012556
Pytorch:https://cloud.tencent.com/developer/article/1521649
---------预训练模型YOLO,DarkNet------------
直接使用pretrained model
网友评论