week01.Ng's CNN Course

CNN 2017.12

1. Computer Vision

2. Edge Detection Example

3. More Edge Detection

• 从明到暗，从暗到明。

• 数字的选择

  $Sobel \ Filter \ldots$

• 可学习的参数

  把这些filter中的参数都当成可学习的，通过backpropagation的方法学习到你想要的filter。

4. Padding

• You do not want your image to shrink.

• The pixels in the corner is used much less than pixels in the center. Throwing away a lot of information near the edge of the image.

• Valid and Same convolution

  $stride = 1$的情况下，
  保证padding后图像大小不变，$n+2p-f+1=n$
  

  $$p = \frac{f-1}{2}$$
  $f$ is odd, symmetric padding; $f$ is even, asymmetric padding.
  通常是odd.

5. Strided convolution

• summary of convolutions

  $n \times n \ image$ $f \times f \ filter$
  $padding \ p$ $stride \ s$
  the size of the output is
  

  $$\lfloor \frac{n+2p-f}{s} + 1 \rfloor \times \lfloor \frac{n+2p-f}{s} + 1 \rfloor$$

• convolution in math textbook

  before convolving rotate the filter by $180^\circ$

6. Convolutions over Volume

• $height \times width \times channels$

• 多个卷积核组合，检测不同的特征

  e.g. vertical edges and horizontal edges

  depth, also refer to the depth of the NN, so we use channels in next videos.

7. One Layer of a Convolutional Network

• notation

8. Simple Convolutional Network Example

• Types of layer in a convolutional network

  convolution
  pooling
  fully connected

9. Pooling Layer

ConvNets often also use pooling layers to reduce the size of the representation, to speed the computation, as well as make some of the features that detects a bit more robust.

• max pooling

So, what the max operates to does is really to say, if these features detected anywhere in this filter, then keep a high number. But if this feature is not detected, so maybe this feature doesn't exist in the upper right-hand quadrant. Then the max of all those numbers is still itself quite small. So maybe that's the intuition behind max pooling. But I have to admit, I think the main reason people use max pooling is because it's been found in a lot of experiments to work well, and the intuition I just described, despite it being often cited, I don't know of anyone fully knows if that is the real underlying reason. I don't have anyone knows if that's the real underlying reason that max pooling works well in ConvNets.

no parameters to learn

• average pooling

  max pooling is more common than vaerage pooling

10. CNN Example

• LeNet-5

11. Why CNN？

• parameter sharing

• sparsity of connections