A brief introduction to CNN
Our general neural network still has shortcomings when it comes to understanding picture information. At this time, convolutional neural network is a booster for computers to process pictures. Convolutional Neural Networks (CNN) is a great tool for neural networks to process picture information. With it, we show pictures to the computer, and the computer understands it more accurately. It is highly recommended to watch the short and concise machine learning-introduction series I made. What is CNN
The leap in computer vision processing has shown strong advantages in image and speech recognition. Before learning convolutional neural networks, we have already assumed that you have a preliminary understanding of neural networks. If not, you can go to the first chapter tensorflow. A video tutorial~
The convolutional neural network includes an input layer, a hidden layer, and an output layer, and the hidden layer includes a convolutional layer and poolinga layer. After the image is input to the convolutional neural network, features are continuously extracted through convolution. Every time a feature is extracted, one will be added feature map. So you will see that the cube in the video tutorial is constantly increasing in thickness, so why is the thickness increasing but getting thinner and thinner? Haha, this is the poolingfunction of the layer, poolingand the layer is downsampling, usually using the maximum value poolingand the average value . value pooling, because there are too many parameters, so poolingthe parameters are sparse so that our network will not be too complicated.
Well, now that you have a general understanding of convolutional neural networks, in the next class we will implement a simple convolutional neural network based on the MNIST dataset through code.