Practical Computer Vision Applications Using Deep Learning with ....pdf

发布时间：2022-05-31 发布人：admin 分类：说明书资料大小：7.10M 资料格式：pdf 举报版权申诉

coolend-10941007-4744300845375998932.pdf-第1页.png

第1页 / 共421页

coolend-10941007-4744300845375998932.pdf-第2页.png

第2页 / 共421页

coolend-10941007-4744300845375998932.pdf-第3页.png

第3页 / 共421页

coolend-10941007-4744300845375998932.pdf-第4页.png

第4页 / 共421页

coolend-10941007-4744300845375998932.pdf-第5页.png

第5页 / 共421页

coolend-10941007-4744300845375998932.pdf-第6页.png

第6页 / 共421页

coolend-10941007-4744300845375998932.pdf-第7页.png

第7页 / 共421页

coolend-10941007-4744300845375998932.pdf-第8页.png

第8页 / 共421页

Table of Contents

About the Author

About the Technical Reviewers

Acknowledgments

Introduction

Chapter 1: Recognition in Computer Vision

Image Recognition Pipeline

Feature Extraction

Color Histogram

Histogram of a Real-World Image

HSV Color Space

GLCM

D Values

θ Values

GLCM Normalization

HOG

Image Gradients

Gradient Direction

Contributing to Histogram Bins

HOG Steps

LBP

Feature Selection & Reduction

Filter

Wrapper

Embedded

Regularization

Chapter 2: Artificial Neural Networks

Introduction to ANNs

Linear Models Are the Base of ANNs

Graphing ANNs

Adjusting Learning Rate for Training ANN

Filter Example

ANN Architecture

Activation Function

Python Implementation

Learning Rate

Testing the Network

Weight Optimization Using Backpropagation

Backpropagation for NN Without Hidden Layer

Weights Update Equation

Why Is the Backpropagation Algorithm Important?

Forward vs. Backward Passes

Partial Derivative

Change in Prediction Error wrt Weights

Prediction Error to Weights Chain

Calculating Chain Partial Derivatives

Interpreting Results of Backpropagation

Updating Weights

Backpropagation for NN with Hidden Layer

Partial Derivatives

Updating Weights

Overfitting

Understand Regularization Based on a Regression Example

Model Capacity/Complexity

L1 Regularization

Designing ANN

Example 1: ANN Without Hidden Layer

Example 2: ANN with a Single Hidden Layer

Chapter 3: Recognition Using ANN with Engineered Features

Fruits 360 Dataset Feature Mining

Feature Mining

Feature Reduction

Filtering Using ANN

ANN Implementation

Engineered Feature Limitations

Not the End of Engineered Features

Chapter 4: ANN Optimization

Introduction to Optimization

Single- vs. Multiobjective Optimization

Best-Parents Selection

Variation Operators

Crossover

Mutation

Python Implementation of an Example

Complete Implementation

NSGA-II

NSGA-II Steps

Dominance

Crowding Distance

Tournament Selection

Crossover

Mutation

Optimizing ANN Using GA

Complete Python Implementation

Chapter 5: Convolutional Neural Networks

From ANN to CNN

The Intuition Behind DL

Derivation of Convolution

Image Analysis Using FC Network

Large Number of Parameters

Neuron Grouping

Pixel Spatial Correlation

Convolution in CNN

Designing a CNN

Pooling Operation for Parameter Reduction

Convolution Operation Example

Max Pooling Operation Example

Building a CNN Using NumPy from Scratch

Reading the Input Image

Preparing Filters

Conv Layer

ReLU Layer

Max Pooling Layer

Stacking Layers

Complete Code

Chapter 6: TensorFlow Recognition Application

Introduction to TF

Tensor

TF Core

Dataflow Graph

Tensor Names

Creating a TF Session

Parameterized Graph Using Placeholder

TF Variables

Variable Initialization

Graph Visualization Using TB

Linear Model

GD Optimizer from TF Train API

Locating Parameters to Optimize

Building FFNN

Linear Classification

Nonlinear Classification

CIFAR10 Recognition Using CNN

Preparing Training Data

Building the CNN

Training CNN

Saving the Trained Model

Complete Code to Build and Train CNN

Preparing Test Data

Testing the Trained CNN Model

Chapter 7: Deploying Pretrained Models

Application Overview

Introduction to Flask

route() Decorator

add_rule_url Method

Variable Rules

Endpoint

HTML Form

File Upload

HTML Inside Flask Application

Flask Templates

Dynamic Templates

Static Files

Deploying Trained Model Using Fruits 360 Dataset

Deploying Trained Model Using CIFAR10 Dataset

Chapter 8: Cross-Platform Data Science Applications

Introduction to Kivy

Basic Application Using BoxLayout

Kivy Application Life Cycle

Widget Size

GridLayout

More Widgets

Widget Tree

Handling Events

KV Language

P4A

Installing Buildozer

Preparing buildozer.spec File

Building Android Application Using Buildozer

Image Recognition on Android

CNN on Android

Appendix A: Installing Your Own Projects Using pip Installer

Creating a Simple Python Project

Project Structure

Project Implementation

Running the Project

Importing the Module into a File Inside Its Directory

Importing the Module into a File Outside Its Directory

How Does Python Locate Libraries?

Manual Installation by Copying Project Files to Site-Packages

How Do Python Installers Locate Libraries?

Preparing the Package and Its Files (__init__.py and setup.py)

__init__.py

setup.py

Distributing the Package

Uploading the Distribution Files Online to Test PyPI

Installing the Distributed Package from Test PyPI

Importing and Using the Installed Package

Using PyPI Rather Than Test PyPI

Index

分享到：

赞收藏

资料库

Practical Computer Vision Applications Using Deep Learning with ....pdf

相关推荐

人工智能

热门标签

最新资料