Convolutional Neural Networks Course Syllabus

Overview: This course provides a comprehensive and practical understanding of Convolutional Neural Networks (CNNs), ideal for those looking to specialize in computer vision. You'll explore foundational concepts and advanced applications through hands-on assignments, guided by renowned AI expert Andrew Ng. The curriculum spans key topics including CNN architecture, object detection, and specialized applications like face recognition and neural style transfer. With approximately 31 hours of content, this course balances theory and implementation, preparing learners to apply CNNs to real-world problems in image classification and beyond.

Module 1: Foundations of Convolutional Neural Networks

Estimated time: 9 hours

• Understand the role of convolutional layers in feature extraction
• Implement pooling layers for dimensionality reduction
• Stack convolutional and pooling layers effectively
• Build CNNs for image classification tasks

Module 2: Deep Convolutional Models: Case Studies

Estimated time: 8 hours

• Study the architecture of ResNets and their residual connections
• Explore the Inception network and its design principles
• Learn practical techniques from seminal research papers
• Apply advanced CNN architectures to complex image tasks

Module 3: Object Detection

Estimated time: 7 hours

• Understand challenges in localizing and detecting multiple objects
• Implement sliding window and convolutional implementations
• Apply the YOLO (You Only Look Once) algorithm for real-time detection

Module 4: Special Applications: Face Recognition & Neural Style Transfer

Estimated time: 7 hours

• Design CNN-based systems for face verification and recognition
• Implement one-shot learning with Siamese networks
• Apply neural style transfer to generate artistic images

Prerequisites

• Familiarity with Python programming
• Basic understanding of machine learning concepts
• Background in linear algebra and calculus recommended

What You'll Be Able to Do After

• Construct and train convolutional neural networks from scratch
• Apply advanced CNN architectures like ResNet and Inception
• Implement object detection systems using YOLO
• Develop face recognition models using deep learning
• Generate artistic images using neural style transfer