011-lec-003. application using cnn
@
LeNet-5
32*32 image, 5*5 filter, 1 stride, 6 filters
$$$\rightarrow$$$1st convolution layer$$$\rightarrow$$$
output will be 28*28*6images
28*28*6images, 2*2 filter, 2 stride
$$$\rightarrow$$$1st pooling$$$\rightarrow$$$
output will be 14*14*6imgaes
14*14*6images, 5*5 filter, 1 stride, 6 filters
$$$\rightarrow$$$2nd convolution$$$\rightarrow$$$
output will be 10*10*16images
16@10*10 image, 2*2 filter, 2 stride
$$$\rightarrow$$$2nd pooling$$$\rightarrow$$$
output will be 5*5*16images
5*5*16images
$$$\rightarrow$$$fully connectioned layer$$$\rightarrow$$$