Building Computer Vision Applications Using Artificial Neural Networks: With Step-by-Step Examples in OpenCV and TensorFlow with Python

Table of Contents
About the Author
About the Technical Reviewer
Acknowledgments
Introduction
Chapter 1: Prerequisites and Software Installation
	Python and PIP
		Installing Python and PIP on Ubuntu
		Installing Python and PIP on macOS
		Installing Python and PIP on CentOS 7
		Installing Python and PIP on Windows
	virtualenv
		Installing and Activating virtualenv
	TensorFlow
		Installing TensorFlow
	PyCharm IDE
		Installing PyCharm
		Configuring PyCharm to Use virtualenv
	OpenCV
		Working with OpenCV
		Installing OpenCV4 with Python Bindings
	Additional Libraries
		Installing SciPy
		Installing Matplotlib
Chapter 2: Core Concepts of Image and Video Processing
	Image Processing
	Image Basics
	Pixels
		Pixel Color
			Grayscale
			Color
	Coordinate Systems
	Python and OpenCV Code to Manipulate Images
		Program: Loading, Exploring, and Showing an Image
		Program: OpenCV Code to Access and Manipulate Pixels
	Drawing
		Drawing a Line on an Image
		Drawing a Rectangle on an Image
		Drawing a Circle on an Image
	Summary
Chapter 3: Techniques of Image Processing
	Transformation
		Resizing
		Translation
		Rotation
		Flipping
		Cropping
	Image Arithmetic and Bitwise Operations
		Addition
		Subtraction
		Bitwise Operations
			AND
			OR
			NOT
			XOR
	Masking
	Splitting and Merging Channels
	Noise Reduction Using Smoothing and Blurring
		Mean Filtering or Averaging
		Gaussian Filtering
		Median Blurring
		Bilateral Blurring
	Binarization with Thresholding
		Simple Thresholding
		Adaptive Thresholding
		Otsu’s Binarization
	Gradients and Edge Detection
		Sobel Derivatives (cv2.Sobel() Function)
		Laplacian Derivatives (cv2.Laplacian() Function)
		Canny Edge Detection
	Contours
		Drawing Contours
	Summary
Chapter 4: Building a Machine Learning–Based Computer Vision System
	Image Processing Pipeline
	Feature Extraction
		How to Represent Features
		Color Histogram
			How to Calculate a Histogram
			Grayscale Histogram
			RGB Color Histogram
		Histogram Equalizer
		GLCM
		HOGs
		LBP
	Feature Selection
		Filter Method
		Wrapper Method
		Embedded Method
	Model Training
		How to Do Machine Learning
		Supervised Learning
		Unsupervised Learning
	Model Deployment
	Summary
Chapter 5: Deep Learning and Artificial Neural Networks
	Introduction to Artificial Neural Networks
		Perceptron
			How a Perceptron Learns
		Multilayer Perceptron
			Why MLP?
		What Is Deep Learning?
		Deep Learning or Multilayer Perceptron Architecture
		Activation Functions
			Linear Activation Function
			Sigmoid or Logistic Activation Function
			TanH/Hyperbolic Tangent
			Rectified Linear Unit
			Leaky ReLU
			Scaled Exponential Linear Unit
			Softplus Activation Function
			Softmax
		Feedforward
		Error Function
			Regression Loss Function
			Binary Classification Loss Function
			Multiclass Classification Loss Function
		Optimization Algorithms
			Gradient Descent
				Local and Global Minima
				Learning Rate
				Regularization
			Stochastic Gradient Descent
				SGD for Distributed and Parallel Computing
				SGD with Momentum
			Adaptive Gradient Algorithm (Adagrad)
			RMSProp
			Adaptive Moment (Adam)
		Backpropagation
	Introduction to TensorFlow
		TensorFlow Installation
		How to Use TensorFlow
		Tensor
		Variable
		Constant
	Our First Computer Vision Model with Deep Learning: Classification of Handwritten Digits
	Model Evaluation
		Overfitting
			Underfitting
			Evaluation Metrics
	Hyperparameters
		TensorBoard
		Experiments for Hyperparameter Tuning
	Saving and Restoring Model
		Save Model Checkpoints During Training
		Manually Save Weights
		Load the Saved Weights and Retrain the Model
		Saving the Entire Model
		Retraining the Existing Model
		Using a Trained Model in Applications
	Convolution Neural Network
		Architecture of CNN
		How Does CNN Work
			Convolution
			Pooling/Subsampling/Downsampling
				Max Pooling
				Average Pooling
		Summary of CNN Concepts
		Training a CNN Model: Pneumonia Detection from Chest X-rays
			Chest X-ray Dataset
			Code Structure
			CNN Model Training
			Pneumonia Prediction
		Examples of Popular CNNs
			LeNet-5
			AlexNet
			VGG-16
		Summary
Chapter 6: Deep Learning in Object Detection
	Object Detection
	Intersection Over Union
	Region-Based Convolutional Neural Network
	Fast R-CNN
	Faster R-CNN
		Region Proposal Network
		Fast R-CNN
	Mask R-CNN
		Backbone
		RPN
		Output Head
		What Is the Significance of the Masks?
		Mask R-CNN in Human Pose Estimation
	Single-Shot Multibox Detection
		SSD Network Architecture
			Multiscale Feature Maps for Detection
				Anchor Boxes and Convolutional Predictors for Detection
			Default Boxes and Aspect Ratios
		Training
			Matching Strategy
			Training Objective
			Choosing Scales and Aspect Ratios for Default Boxes
			Hard Negative Mining
			Data Augmentation
			Nonmaximum Suppression
		SSD Results
	YOLO
		YOLO Network Design
	Limitations of YOLO
		YOLO9000 or YOLOv2
		YOLOv3
	Comparison of Object Detection Algorithms
		Comparison of Architecture
		Comparison of Performance
	Training Object Detection Model Using TensorFlow
		TensorFlow on Google Colab with GPU
			Accessing Google Colab
			Connecting to the Hosted Runtime
			Selecting a GPU Hardware Accelerator
			Creating a Colab Project
			Setting the Runtime Environment for TensorFlow and Model Training
			Downloading the Oxford-IIIT Pet Dataset
			Generating TensorFlow TFRecord Files
			Downloading a Pre-trained Model for Transfer Learning
			Configuring the Object Detection Pipeline
			Executing the Model Training
			Exporting the TensorFlow Graph
			Downloading the Object Detection Model
			Visualizing the Training Result in TensorBoard
	Detecting Objects Using Trained Models
		Installing TensorFlow’s models Project
		Code for Object Detection
	Training a YOLOv3 Model for Object Detection
		Installing the Darknet Framework
		Downloading Pre-trained Convolutional Weights
		Downloading an Annotated Oxford-IIIT Pet Dataset
		Preparing the Dataset
		Configuring the Training Input
		Configuring the Darknet Neural Network
		Training a YOLOv3 Model
		How Long the Training Should Run
		Final Model
	Detecting Objects Using a Trained YOLOv3 Model
		Installing Darknet to the Local Computer
		Python Code for Object Detection
	Summary
Chapter 7: Practical Example: Object Tracking in Videos
	Preparing the Working Environment
	Reading a Video Stream
	Loading the Object Detection Model
	Detecting Objects in Video Frames
	Creating a Unique Identity for Objects Using dHash
	Using the Hamming Distance to Determine Image Similarity
	Object Tracking
	Displaying a Live Video Stream in a Web Browser
		Installing Flask
		Flask Directory Structure
		HTML for Displaying a Video Stream
		Flask to Load the HTML Page
		Flask to Serve the Video Stream
		Running the Flask Server
	Putting It All Together
	Summary
Chapter 8: Practical Example: Face Recognition
	FaceNet
		FaceNet Neural Network Architecture
			Input Images
			Deep CNN
			Face Embedding
			Triplet Loss Function
			Triplet Selection
	Training a Face Recognition Model
		Checking Out FaceNet from GitHub
		Dataset
		Downloading VGGFace2 Data
		Data Preparation
		Model Training
		Evaluation
	Developing a Real-Time Face Recognition System
		Face Detection Model
		Classifier for Face Recognition
			Face Alignment
			Classifier Training
			Face Recognition in a Video Stream
	Summary
Chapter 9: Industrial Application: Real-Time Defect Detection in Industrial Manufacturing
	Real-Time Surface Defect Detection System
		Dataset
		Google Colab Notebook
		Data Transformation
		Training the SSD Model
		Exporting the Model
		Model Evaluation
		Prediction
		Real-Time Defect Detector
	Image Annotations
		Installing VoTT
		Create Connections
		Create a New Project
		Create Class Labels
		Label the Images
		Export Labels
	Summary
Chapter 10: Computer Vision Modeling on the Cloud
	TensorFlow Distributed Training
		What Is Distributed Training?
			Data Parallelism
			Model Parallelism
		TensorFlow Distribution Strategy
			MirroredStrategy
			CentralStorageStrategy
			MultiWorkerMirroredStrategy
				Cluster Configuration
				Dataset Sharding
				Fault Tolerance
			TPUStrategy
			ParameterServerStrategy
			OneDeviceStrategy
		TF_CONFIG: TensorFlow Cluster Configuration
			An Example TF_CONFIG
	Example Code of Distributed Training with a Parameter Server
	Steps for Running Distributed Training on the Cloud
	Distributed Training on Google Cloud
		Signing Up for GCP Access
		Creating a Google Cloud Storage Bucket
		Creating the GCS Bucket from the Web UI
		Creating the GCS Bucket from the Cloud Shell
		Launching GCP Virtual Machines
		SSH to Log In to Each VMs
		Uploading the Code for Distributed Training or Cloning the GitHub Repository
		Installing Prerequisites and TensorFlow
		Running Distributed Training
	Distributed Training on Azure
		Creating a VM with Multiple GPUs on Azure
		Installing GPU Drivers and Libraries
		Creating virtualenv and Installing TensorFlow
		Implementing MirroredStrategy
		Running Distributed Training
	Distributed Training on AWS
		Horovod
		How to Use Horovod
		Creating a Horovod Cluster on AWS
			Horovod Cluster
			Running Distributed Training
		Installing Horovod
		Running Horovod to Execute Distributed Training
	Summary
Index