day 34

CNN

Convolution - > refers to finding or extracting features
CNN - > in real is CANN - > Convolution Artificial Neural Network
Convolution is divided into three parts
1. Convolution - > extract feature from image,video or anything
2. Max pulling -> extracting maximum number value cam in feature map by matching the minimum matrix of 2x2
3. Flattening - > converting the max pulling into a vector and is finally used for ANN
Feature map - > an sample image that we wish to find in a particular image is matched and checked over the 2nd image and the resultant output is stored in another table called feature map
The size of sample image depends upon the CPU and GPU of system
Overfitting - > problem that arises by training the model again and again and causes a diminishing in accuracy, due to possible occurence of impurity in data ( chaning minimum data )
Workon and Study everything on imageai.org
ANN prefers cluster computing
pjreddie.net/darknet/yolo