Convolutional-Neural-Network-Part1

Convolutional Neural Network 개요

	DMLP(Deep Multi Layer Perceptron)	Convolutional Neural Network
network structure	Fully-Connected	Partially-Connected
Learning speed	Very Low	Low
장점		Model Complexity를 획기적으로 낮춤.
단점	Over-fitting 가능성	모든 feature추출을 해야 partially-connected 단점 극복가능함.
사용처		Image Processing, Signal Processing