Deep Learning Complete Course | Part 2| CNN implementation.

Instructor – Akarsh Vyas
Welcome back! In this video, we’ll take the next big step in Deep Learning and dive deep into Convolutional Neural Networks (CNNs) the architecture that completely changed the world of Computer Vision.

You can download the code and datasets from here:
Code Link – https://github.com/AkarshVyas/CNN-video

📘 All the notes of our classes are here:
Notes – https://drive.google.com/file/d/15b8U3Zo3WO-v9J93Is_h2ce7WVfC5ar9/view?usp=drive_link

Here’s what you’ll learn in this CNN deep dive:

The problem with ANN on images and why CNN was invented

The intuition behind Convolutions, Filters, and Feature Maps

Pooling layers and why they make CNNs efficient

Step-by-step architecture: Convolution → Pooling → Fully Connected

Famous CNN models (LeNet, AlexNet, VGG, ResNet) and how they shaped modern AI

Hands-on coding: Building CNNs with TensorFlow/Keras on real datasets

Applications of CNNs in real life — from face recognition to self-driving cars

These are the most important building blocks of modern Computer Vision. If you’ve understood ANNs from our first video, this session will complete the foundation you need before moving to advanced architectures and real-world AI projects.

By the end of this video, you’ll not only understand how CNNs work but also be able to build and train your own CNN from scratch.

0:00:00 – 00:00:38 – introduction
00:00:38 – 00:11:53- CNN(introduction)
00:11:55 – 00:19:00 – How CNN works
00:19:00 – 00:21:34 – Understanding the Architecture
00:21:34 – 00:25:03 – Layers in CNN
00:25:03 – 00:42:12 – Edge Finding in CNN
00:42:12 – 00:48:52 – understanding(padding and strides)
00:48:52 – 00:55:26 – why we use strides
00:55:26 – 01:04:41- pooling
01:04:41 – 01:06:47- max pooling
01:06:47 – 01:15:20 – Revision and flattening
01:15:20 – 01:54:13 – code implementation
01:54:13 – 01:54:47 – outro

source