目录:git
0 计算机视觉四大基本任务0 计算机视觉四大基本任务
https://zhuanlan.zhihu.com/p/31727402github
目标检测解读汇总算法
https://zhuanlan.zhihu.com/p/36402362网络
图像语义分割综述ide
https://zhuanlan.zhihu.com/p/37801090函数
1 经典卷积网络
AlexNet学习
VGG优化
GoogleNetorm
ResNet:blog
Xception
DenseNet
shufflenet:https://zhuanlan.zhihu.com/p/48261931
2 卷积、空洞卷积
卷积:https://www.zhihu.com/question/22298352
空洞卷积:https://www.zhihu.com/question/54149221/answer/323880412
https://zhuanlan.zhihu.com/p/50369448
3 正则化
batch normalization、group normalization:
https://zhuanlan.zhihu.com/p/56219719
L一、L2:
4 全卷积网络
5 1*1卷积核
https://zhuanlan.zhihu.com/p/40050371
https://www.zhihu.com/question/56024942
6 感觉野
https://zhuanlan.zhihu.com/p/28492837
7 常见损失
损失函数改进方法总览:
focal loss:https://zhuanlan.zhihu.com/p/32423092
交叉熵:
对比损失(Contrastive Loss):
三元组损失(Triplet Loss):
8 优化算法
Momentum:https://zhuanlan.zhihu.com/p/21486826
Nesterov Accelerated Gradient:https://zhuanlan.zhihu.com/p/22810533
9 concat 和 add的区别
https://www.zhihu.com/question/306213462/answer/562776112
10 注意力机制
https://zhuanlan.zhihu.com/p/32928645
https://zhuanlan.zhihu.com/p/56501461
11 CNN(卷积神经网络)、RNN(循环神经网络)、DNN(深度神经网络)
https://www.zhihu.com/question/34681168/answer/84061846
12 边框回归
13 非极大值回归(NMS, Soft NMS)
Soft NMS:
14 激活函数
https://zhuanlan.zhihu.com/p/22142013
https://www.zhihu.com/question/59031444/answer/177786603
https://www.zhihu.com/question/29021768/answer/43488153
15 评价指标
目标检测mAP
https://github.com/rafaelpadilla/Object-Detection-Metrics
语义分割(PA、MPA、MIoU、FWIoU)
https://zhuanlan.zhihu.com/p/61880018
https://link.zhihu.com/?target=https%3A//arxiv.org/abs/1704.06857
16 batch size的选择
https://www.zhihu.com/question/61607442/answer/440401209
https://www.zhihu.com/question/32673260/answer/71137399
17 Graph Convolutional Network(GCN)
https://www.zhihu.com/question/54504471/answer/332657604