Convolution Neural Network, CNN

1. Overview of the types of deep learning

Models, types of learning algorithms	Major research area
Convolutional Neural Network	video perception, computer vision
Recurrent Neural Network	voice perception, weather/stock price prediction
Restricted Boltzmann Machine	classification, 회귀 분석
Deep Belief Network	writing, voice perception
Generative Adevrsarial Network (생산적 적대 신경망 )	음성, 영상 복원, 생성, 변조

 

2. Meaning of the Convolution 

  1. Convolution : Operation used in processing the math&signal 

3. Convolution Neural Network, CNN

  1. Used actively in deep neural network of the feild of the video perception

  2. Meaning of the convolution in the video perception is "adding and multipling each value by filter value"

  3. Each weight of each fillter is updated during backpropagation along with the connection strength between nodes 

 

4. Example of the Convolution Neural Network

  - Extracting the feature of each part of video and then, video perception, classification by convey to the Deep Neural Network

  - Unlike exiting filtering techniques, the optimization of Convoution filter  weights is done using by Backpropagation