Computer Vision Using Deep Learning: Neural Network Architectures with Python and Keras

✍ Scribed by Vaibhav Verdhan

• Publisher: Apress
• Year: 2021
• Tongue: English
• Leaves: 320
• Category: Library

✦ Synopsis

Organizations spend huge resources in developing software that can perform the way a human does. Image classification, object detection and tracking, pose estimation, facial recognition, and sentiment estimation all play a major role in solving computer vision problems.

This book will bring into focus these and other deep learning architectures and techniques to help you create solutions using Keras and the TensorFlow library. You'll also review mutliple neural network architectures, including LeNet, AlexNet, VGG, Inception, R-CNN, Fast R-CNN, Faster R-CNN, Mask R-CNN, YOLO, and SqueezeNet and see how they work alongside Python code via best practices, tips, tricks, shortcuts, and pitfalls. All code snippets will be broken down and discussed thoroughly so you can implement the same principles in your respective environments.

Computer Vision Using Deep Learning offers a comprehensive yet succinct guide that stitches DL and CV together to automate operations, reduce human intervention, increase capability, and cut the costs.

What You'll Learn

• Examine deep learning code and concepts to apply guiding principals to your own projects
• Classify and evaluate various architectures to better understand your options in various use cases
• Go behind the scenes of basic deep learning functions to find out how they work

Who This Book Is For

Professional practitioners working in the fields of software engineering and data science. A working knowledge of Python is strongly recommended. Students and innovators working on advanced degrees in areas related to computer vision and Deep Learning.

✦ Table of Contents

Table of Contents
About the Author
About the Technical Reviewer
Acknowledgments
Introduction
Chapter 1: Introduction to Computer Vision and Deep Learning
1.1 Technical requirements
1.2 Image Processing using OpenCV
1.2.1 Color detection using OpenCV
1.3 Shape detection using OpenCV
1.3.1 Face detection using OpenCV
1.4 Fundamentals of Deep Learning
1.4.1 The motivation behind Neural Network
1.4.2 Layers in a Neural Network
1.4.3 Neuron
1.4.4 Hyperparameters
1.4.5 Connections and weight of ANN
1.4.6 Bias term
1.4.7 Activation functions
1.4.7.1 Sigmoid function
1.4.7.2 tanh function
1.4.7.3 Rectified Linear Unit or ReLU
1.4.7.4 Softmax function
1.4.8 Learning rate
1.4.9 Backpropagation
1.4.10 Overfitting
1.4.11 Gradient descent
1.4.12 Loss functions
1.5 How Deep Learning works?
1.5.1 Popular Deep Learning libraries
1.6 Summary
1.6.1 Further readings
Chapter 2: Nuts and Bolts of Deep Learning for Computer Vision
2.1 Technical requirements
2.2 Deep Learning using TensorFlow and Keras
2.3 What is a tensor?
2.3.1 What is a Convolutional Neural Network?
2.3.2 What is convolution?
2.3.3 What is a Pooling Layer?
2.3.4 What is a Fully Connected Layer?
2.4 Developing a DL solution using CNN
2.5 Summary
2.5.1 Further readings
Chapter 3: Image Classification Using LeNet
3.1 Technical requirements
3.2 Deep Learning architectures
3.3 LeNet architecture
3.4 LeNet-1 architecture
3.5 LeNet-4 architecture
3.6 LeNet-5 architecture
3.7 Boosted LeNet-4 architecture
3.8 Creating image classification models using LeNet
3.9 MNIST classification using LeNet
3.10 German traffic sign identification using LeNet
3.11 Summary
3.11.1 Further readings
Chapter 4: VGGNet and AlexNet Networks
4.1 Technical requirements
4.2 AlexNet and VGG Neural Networks
4.3 What is AlexNet Neural Network?
4.4 What is VGG Neural Network?
4.5 VGG16 architecture
4.6 Difference between VGG16 and VGG19
4.7 Developing solutions using AlexNet and VGG
4.8 Working on CIFAR-10 using AlexNet
4.9 Working on CIFAR-10 using VGG
4.10 Comparing AlexNet and VGG
4.11 Working with CIFAR-100
4.12 Summary
4.12.1 Further readings
Chapter 5: Object Detection Using Deep Learning
5.1 Technical requirements
5.2 Object Detection
5.2.1 Object classification vs. object localization vs. object detection
5.2.2 Use cases of Object Detection
5.3 Object Detection methods
5.4 Deep Learning frameworks for Object Detection
5.4.1 Sliding window approach for Object Detection
5.5 Bounding box approach
5.6 Intersection over Union (IoU)
5.7 Non-max suppression
5.8 Anchor boxes
5.9 Deep Learning architectures
5.9.1 Region-based CNN (R-CNN)
5.10 Fast R-CNN
5.11 Faster R-CNN
5.12 You Only Look Once (YOLO)
5.12.1 Salient features of YOLO
5.12.2 Loss function in YOLO
5.12.3 YOLO architecture
5.13 Single Shot MultiBox Detector (SSD)
5.14 Transfer Learning
5.15 Python implementation
5.16 Summary
5.16.1 Further readings
Chapter 6: Face Recognition and Gesture Recognition
6.1 Technical toolkit
6.2 Face recognition
6.2.1 Applications of face recognition
6.2.2 Process of face recognition
6.2.2.1 Deep Learning modes for face recognition
6.2.3 DeepFace solution by Facebook
6.2.4 FaceNet for face recognition
6.2.5 Python implementation using FaceNet
6.2.6 Python solution for gesture recognition
6.3 Summary
6.3.1 Further readings
Chapter 7: Video Analytics Using Deep Learning
7.1 Technical toolkit
7.2 Video processing
7.3 Use cases of video analytics
7.4 Vanishing gradient and exploding gradient problem
7.5 ResNet architecture
7.5.1 ResNet and skip connection
7.5.2 Inception network
7.5.2.1 1x1 convolutions
7.5.3 GoogLeNet architecture
7.5.4 Improvements in Inception v2
7.6 Video analytics
7.7 Python solution using ResNet and Inception v3
7.8 Summary
7.8.1 Further readings
Chapter 8: End-to-End Model Development
8.1 Technical requirements
8.2 Deep Learning project requirements
8.3 Deep Learning project process
8.4 Business problem definition
8.4.1 Face detection for surveillance
8.4.1.1 Defect detection for manufacturing
8.4.2 Source data or data discovery phase
8.4.2.1 Face detection for identification
8.4.2.2 Live environment on a manufacturing line
8.5 Data ingestion or data management
8.6 Data preparation and augmentation
8.6.1 Image augmentation
8.7 Deep Learning modeling process
8.7.1 Transfer learning
8.7.2 Common mistakes/challenges and boosting performance
8.8 Model deployment and maintenance
8.9 Summary
8.9.1 Further readings
References
Major activation functions and layers used in CNN
Google Colab
Index

✦ Subjects

Computers, Intelligence (AI) & Semantics, Information Technology