fast.ai reference
fast.aicoursematerialbook
Search…
Fast.ai reference
How to run the code
PyTorch
Jupyter Notebook
Computer Vision
Convolution
Pooling
Tabular data
NLP
Reinforcement Learning
Overfitting and Underfitting
Contribute
FEEDBACK & CORRECTION
Powered By GitBook
Convolution
If we represent an image as a two dimensional array of pixels, then convolutions can be applied to detect edges and other more complex features.
Example greyscale image
The image is then transformed from an array of pixels to an array of convolutions, which are then again fed into more and more convolutional layers. This enables the learning of complex shapes that can then be used by the final fully connected layer to classify the image.
Intuitively, one can imagine the process going like this:
  1. 1.
    An image of a car is received
  2. 2.
    Features are extracted with convolutional layers
  3. 3.
    The resulting vector of features is then classified
    i.e. this object has four wheels, windows, a license plate and seat --> it must be a car.

Convolutions Mathematically

In mathematical terms, in a convolution a small kernel (usually 2x2, 3x3 or 4x4) is applied to an input matrix (the pixels of an image).
A 3x3 kernel is applied to an 5x5 image
The 3x3 kernel is applied to every 3x3 square in the input image sequentially, reducing the image dimension and extracting information in the process.
Previous
Computer Vision
Next
Pooling
Last modified 2yr ago
Export as PDF
Copy link