Convolutional Neural Network #ai #education #cnn #neuralnetworks #deeplearning #machinelearning

Added: 2026-05-17

[00:00:00]What is a convolutional neural network and how does it extract features from an image? So, CNN is a type of deep learning model designed to process and analyze visual data. Now, here you can see an image, but for a computer, it's just a giant grid of numbers representing colors. Now, let's understand how CNN actually works. So, first the CNN uses a filter which captures one specific feature and we have multiple features to capture all the features from an image. So, the filters will slide through the grid and extract the features and store it in M cross N matrix. But, if we use filter at the edges, then there is a high chance that it might not capture the details in the corners. So, for that, we apply padding to extend the image so that the dot product or the feature extraction will be done more accurately. Then, it is passed through an activation function which throws away any information that doesn't fit and just keeps useful parts.

[00:00:52]And after that, pooling is done which turns the detail map into a much smaller summary by just keeping most important features so that the image will get even much more smaller and computation will be easier. And after getting this, we flatten this matrix and pass it to an artificial neural network. Now, if you want to learn about transformer architecture, then subscribe me to watch my next video on this.