section>
The correct bibliographic citation for this manual is as follows: Blanchard, Robert 2020. Deep Learning for Computer Vision with SAS®: An Introduction. Cary, NC: SAS Institute Inc.
Deep Learning for Computer Vision with SAS®: An Introduction
Copyright © 2020, SAS Institute Inc., Cary, NC, USA
ISBN 978-1-64295-972-7 (Hardcover)
ISBN 978-1-64295-915-4 (Paperback)
ISBN 978-1-64295-916-1 (PDF)
ISBN 978-1-64295-917-8 (EPUB)
ISBN 978-1-64295-918-5 (Kindle)
All Rights Reserved. Produced in the United States of America.
For a hard copy book: No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, or otherwise, without the prior written permission of the publisher, SAS Institute Inc.
For a web download or e-book: Your use of this publication shall be governed by the terms established by the vendor at the time you acquire this publication.
The scanning, uploading, and distribution of this book via the Internet or any other means without the permission of the publisher is illegal and punishable by law. Please purchase only authorized electronic editions and do not participate in or encourage electronic piracy of copyrighted materials. Your support of others’ rights is appreciated.
U.S. Government License Rights; Restricted Rights: The Software and its documentation is commercial computer software developed at private expense and is provided with RESTRICTED RIGHTS to the United States Government. Use, duplication, or disclosure of the Software by the United States Government is subject to the license terms of this Agreement pursuant to, as applicable, FAR 12.212, DFAR 227.7202-1(a), DFAR 227.7202-3(a), and DFAR 227.7202-4, and, to the extent required under U.S. federal law, the minimum restricted rights as set out in FAR 52.227-19 (DEC 2007). If FAR 52.227-19 is applicable, this provision serves as notice under clause (c) thereof and no other notice is required to be affixed to the Software or documentation. The Government’s rights in Software and documentation shall be only those set forth in this Agreement.
SAS Institute Inc., SAS Campus Drive, Cary, NC 27513-2414
June 2020
SAS® and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA and other countries. ® indicates USA registration.
Other brand and product names are trademarks of their respective companies.
SAS software may be provided with certain third-party software, including but not limited to open-source software, which is licensed under its applicable third-party software license agreement. For license information about third-party software distributed with SAS software, refer to http://support.sas.com/thirdpartylicenses.
Contents
Chapter 1: Introduction to Deep Learning
Introduction to Neural Networks
Traditional Neural Networks versus Deep Learning
Building a Deep Neural Network
Demonstration 1: Loading and Modeling Data with Traditional Neural Network Methods
Demonstration 2: Building and Training Deep Learning Neural Networks Using CASL Code
Chapter 2: Convolutional Neural Networks
Introduction to Convoluted Neural Networks
Input Layers
Convolutional Layers
Using Filters
Padding
Feature Map Dimensions
Pooling Layers
Traditional Layers
Demonstration 1: Loading and Preparing Image Data
Demonstration 2: Building and Training a Convolutional Neural Network
Introduction
Architectural Design Strategies
Image Preprocessing and Data Enrichment
Transfer Learning Introduction
Domains and Subdomains
Types of Transfer Learning
Transfer Learning Biases
Transfer Learning Strategies
Customizations with FCMP
Tuning a Deep Learning Model
Introduction
Types of Object Detection Algorithms
Data Preparation and Prediction Overview
Normalized Locations
Multi-Loss Error Function
Error Function Scalars
Anchor Boxes
Final Convolution Layer
Demonstration: Using DLPy to Access SAS Deep Learning Technologies: Part 1
Demonstration: Using DLPy to Access SAS Deep Learning Technologies: Part 2
Chapter 5: Computer Vision Case Study
About This Book
What Does This Book Cover?
Deep learning is an area of machine learning that has become ubiquitous with artificial intelligence. The complex, brain-like structure of deep learning models is used to find intricate patterns in large volumes of data. These models have heavily improved the performance of general supervised models, time series, speech recognition, object detection and classification, and sentiment analysis.
SAS has a rich