How nature inspired digital eyes
The math of sliding windows
The great ImageNet breakthrough
Computer vision in the wild
YOUR GOAL
Master of How computers learned to see the world
007skaty CNN Sliding Window Math
Subtract the filter size from the padded space, And divide by the stride to find the output's place.
007skaty Max pooling scans two-by-two pixel grids, keeping only the highest value to cut the overall image size in
007skaty A color-coded vector diagram shows a sliding filter grid creating a feature map
007skaty Slide the window across the pixel grid to discover how AI processes images.
Pixel math: To a computer, a black-and-white image is just a grid of numbers from zero (pure black) to 255 (pure white).
A long, decisive step. In 1998, Yann LeCun's LeNet-5 system used a stride of one to scan handwritten zip codes.
Which layer in a convolutional neural network actually makes the final decision, like 'this is a cat'?
What is the sliding grid of numbers used to detect features in a convolutional neural network called?
What do the very first layers of a deep convolutional neural network typically look for?