Table of Contents
show
Computer Vision
One of the applications of Deep Learning
Examples: Image Classification, Object Detection, Neuro style transfer
Challenges of computer vision
Input size can be too big
Deep Learning for Larger images
- In the input layer, number of features will be 3 million
- In the first hidden layer, there are 1000 units, the weight for w[1] : (1000, 3M)
- i.e., the matix will be having 3 billion parameters which is very very large
- With the number of parameters, it is difficult to get enough data to prevent a neural network from overfitting
What is an edge?
Also, the computational requirements and memory requirements become infeasible for training large neural network
Summary
In computer vision applications, the size of the image may be very high
- To process with that, convolution operation may be required
- Convolution is one of the fundamental building blocks of convolutional neural network
Views: 3