[PART 0] Prologue

0.1 What is deep learning image analysis?
0.2 What can you do with an AI camera?
0.3 Preparing for the Practice of This Book
___Preparing for hands-on training in Google Colab
___Starting the Windows Project Environment

[PART 1] Let's learn deep learning video analysis.

▣ Chapter 1: Introduction to Deep Learning Image Analysis
1.1 Three representative deep learning image analysis techniques
___Image Classification
___Image Object Recognition
___image segmentation
1.2 Understanding Deep Learning for Image Analysis
Deep Learning in the History of AI
Understanding Deep Neural Networks
1.3 Basic Structure of a Deep Learning Model
___Deep Learning Training Process and Inference
___Optimization of loss function and weights
___Gradient descent and backpropagation
___Softmax function
1.4 ANN MNIST PyTorch Example

▣ Chapter 2: The Beginning of Deep Learning Image Analysis, CNN
2.1 Why are CNNs important in deep learning image analysis?
___Characteristics of input data in image analysis
___FC layer and Conv layer
2.2 Understanding CNN
___Difference between activation maps and feature maps
___The role of Conv and Pooling layers
2.3 Preparing for the Deep Learning Training Process
___Prepare the dataset
___Activation function
___LeNet: CNN MNIST PyTorch Example

▣ Chapter 3: Learning Process for Deep Learning Image Analysis
3.1 Weight Optimization Solvers
___SGD + Momentum
___Adagrad
___RMSProp
___Adam
3.2 How to improve deep learning results
___Batch Normalization
___Data Augmentation and Transfer Learning
3.3 Popular CNN network architectures
___AlexNet: Winner of the First CNN-Based Image Classification Competition
___VGGNet: A Simple, High-Performance Network
___GoogLeNet: A network created by Google and used by everyone
___ResNet: The Deepest and Highest-Performing Network
3.4 ResNet PyTorch Example
___Training a PyTorch classification model
___PyTorch Classification Model Inference

▣ Chapter 4: Image Segmentation and Object Recognition
4.1 Image Segmentation
___Image segmentation concept
Evaluation criteria for the ___ model
___FCN image segmentation
4.2 Image Object Recognition
___Basic concepts of image object recognition
___Faster R-CNN
4.3 YOLO: The First Real-Time Object Recognition Network
___YOLO: You Only Look Once
___YOLOv2: Better, Faster, More Powerful
___YOLOv3: Incremental Improvement
4.4 SSD: Combining the strengths of Faster R-CNN and YOLO
___SSD: Single Shot Multi Box Detector
4.5 Other networks
___Mask R-CNN: Image Object Segmentation
___MobileNet v2: Small but Powerful Object Recognition
___YOLOv4: The new YOLO
___YOLOv4-tiny: Tiny version for small devices
4.6 YOLOv4 Practice
___Preparing for the practice
___learning
___Deep Neural Network Learning
___Testing Image Inference in Colab

[PART 2] Video Analysis Project Using Deep Learning

▣ Chapter 5: Classifying Recyclables Using Image Classification
5.1 Overview of the Recyclable Waste Separation Project
5.2 Dataset Class
___PyTorch Custom Dataset Class
___Dataset class for image classification
5.3 Implementing a Deep Neural Network
Import ___ module
___MODEL's core structure
Implementing the __init__ method
Implementing the ___forward method
5.4 Implementing a Transfer Learning Deep Neural Network
5.5 Deep Neural Network Learning Class
Structure of the ___ class
Implementing the __init__ method
Implementing the ___prepare_network method
Implementing the ___training_network method
5.6 Training with Deep Neural Networks in Colab
___Download Python classes and datasets
___learning
Download the ___model
5.7 Testing Image Inference in Colab
5.8 Inference Practice in a Windows Environment
___Recyclable Classification Inference Code
___Running Recyclables Classification Inference
5.9 Summary

▣ Chapter 6: SSD Object Recognition CCTV
6.1 Project Goals and User Scenarios
6.2 Preparing for the Lab
___Configure Google Drive and download the source code
___Download the Open Image Dataset
___Dataset class (OpenImagesDataset)
6.3 Network Learning
___SSD MobileNet v1 deep neural network
___SSD MobileNet v1 training class
___Training SSD MobileNet v1
___Testing Inference in Colab
6.4 Project Inference Practice in a Windows Environment
___SSD Object Recognition CCTV Inference (Video and Camera Footage Inference) Code Description
___SSD Object Recognition CCTV Inference Application (Object Detection Logging) Code Description
___SSD Object Recognition CCTV Inference Execution
6.5 Summary

▣ Chapter 7: Crosswalk Pedestrian Protection System Using YOLO
7.1 Project Goals and User Scenarios
7.2 Preparing for YOLOv5 Practice
___Prepare the dataset
___YOLOv5 Training
7.3 YOLOv5 Inference Practice on Windows
___Preparing YOLOv5 Inference on Windows
___YOLOv5 Inference Test
7.4 Preparing for YOLOv7 Practice
___YOLOv7 Description
___Prepare the dataset
___YOLOv7 Training
YOLOv7 Inference Practice on Windows 7.5
___Preparing YOLOv7 inference on Windows
___YOLOv7 Inference Test
7.6 Summary

[PART 3] Using Jetson Nano and JetBot

▣ Chapter 8: Jetson Nano Inference Project
8.1 Getting Started with Jetson Nano
___Jetson Nano Specifications
___Booting the Jetson Nano
___Check the basic GitHub source
___Jetson Nano Camera Test
___Check the Jetson Nano deep learning basic settings
___AI Acceleration Engine
8.2 Recyclable Classification Inference on Jetson Nano
___Preparing to classify recyclables
___Running Recyclables Classification Inference
8.3 SSD Object Recognition CCTV Jetson Nano Inference
___SSD inference test of still photos
___Video and Camera Image Inference Code
___SSD Object Recognition CCTV Inference Application (Object Detection Logging) Code Description
___SSD Object Recognition CCTV Inference Execution
___SSD Object Recognition CCTV Inference Application (Object Detection Logging) Execution
8.4 YOLOv4 Field Filming Black Box Jetson Nano Inference
___Prepare for inference
___Darknet Black Box Inference Code
8.5 YOLOv5 Crosswalk Pedestrian Protection System Jetson Nano Inference
___Preparing for Inference on Jetson Nano
___TensorRT Optimization
Inference using ___TensorRT
8.6 Summary

▣ Chapter 9: Three AI Mobile Robot Projects
9.1 Preparing for the AI ​​Mobile Robot Project
___AI mobile robot preparation
___AI Mobile Robot Basic Environment
___AI Mobile Robot Basic Test
___Power Mode Settings
9.2 Automatic Emergency Braking for Collision Avoidance
___Collecting situational data for collision avoidance
___Training with the AlexNet network
___Automatic Emergency Braking Code for Collision Avoidance
___AI-mobile robot automatic emergency braking test
9.3 Following Bot
___Inference Acceleration Engine Library
___Inference Acceleration Object Recognition Model Example Code
___Followingbot test code
___Testing a following bot with a pre-trained model
Testing a Following Bot with an SSD model for CCTV
9.4 Lane-Aware Autonomous Driving
___Lane Data Collection (Simple Test)
___Training a regression model with the ResNet-18 network
___Lane Recognition Autonomous Driving Test (Simple Test)
___Lane Recognition Autonomous Driving Test Run
___Lane data collection and training (track testing)
___Lane Recognition Autonomous Driving Test (Track Test)
9.5 Summary

★ What this book covers ★

◎ Core Theory of Deep Learning Image Processing
◎ Famous CNN architecture and SSD for object recognition, YOLOv4~v11
◎ Image classification and object recognition practice
◎ Computer vision project using NVIDIA Jetson Nano
◎ Mobile robot (JetBot) autonomous driving project

★ YouTube channel to help you practice this book ★

https://www.youtube.com/@r-ssaem