[딥러닝] Image Classification Models and Dataset

1. Image Classification Model

1.1. LeNet5 (Yann LeCun, 1998)

• First CNN network
• MLP network 단점 극복
• 2D input [32*32]
• LeNet5 in Lecutre Andrew Ng - C4W2L02 Classic Network 0:11

1.2. AlexNet (Alex Krizhevsky, 2012)

• Input [224*224]
• Multi-GPU training
• ReLU
• AlexNet in Lecutre Andrew Ng - C4W2L02 Classic Network 7:02

1.3. VGGNet (Karen Simonyan, Andrew Zisserman, 2014)

• 3x3 convolution
• Deep layers
• VGG-16 in Lecutre Andrew Ng - C4W2L02 Classic Network 12:35

1.4. GoogLeNet (Christian Szegedy, 2014)

• Inception module
• 1x1 convolution
• GoogLeNet in Lecutre
  • Andrew Ng - C4W2L05 Network In Network
  • Andrew Ng - C4W2L06 Inception Network Motivation
  • Andrew Ng - C4W2L07 Inception Network
• zzwon1212 - GoogLeNet (1 x 1 convolution and Inception network)

1.5. ResNet (Kaiming He, 2015)

• Residual block
• Skip connection
• ResNet in Lecutre
  • Andrew Ng - C4W2L03 Resnets
  • Andrew Ng - C4W2L04 Why ResNets Work
• zzwon1212 - ResNet (Residual block and Skip connection)
  

2. Benchmarks & Public dataset

---

📙 참고

• LeNet5, AlexNet, VGG-16
  • Andrew Ng - C4W2L02 Classic Network
• GoogLeNet
  • Andrew Ng - C4W2L05 Network In Network
  • Andrew Ng - C4W2L06 Inception Network Motivation
  • Andrew Ng - C4W2L07 Inception Network
  • zzwon1212 - 1 x 1 Convolution and Inception Network
• ResNet
  • Andrew Ng - C4W2L03 Resnets
  • Andrew Ng - C4W2L04 Why ResNets Work