Machine Learning for Tomographic Imaging. Professor Ge Wang

      

      Machine Learning for

      Tomographic Imaging

       Ge Wang

       Rensselaer Polytechnic Institute

       Yi Zhang

       Sichuan University

       Xiaojing Ye

       Georgia State University

       Xuanqin Mou

       Xi’an Jiaotong University

      IOP Publishing, Bristol, UK

      Copyright © IOP Publishing Ltd 2020

      All rights reserved. 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, recording or otherwise, without the prior permission of the publisher, or as expressly permitted by law or under terms agreed with the appropriate rights organization. Multiple copying is permitted in accordance with the terms of licences issued by the Copyright Licensing Agency, the Copyright Clearance Centre and other reproduction rights organizations.

      Permission to make use of IOP Publishing content other than as set out above may be sought at [email protected].

      Ge Wang, Yi Zhang, Xiaojing Ye and Xuanqin Mou have asserted their right to be identified as the authors of this work in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988.

      ISBN 978-0-7503-2216-4 (ebook)

      ISBN 978-0-7503-2214-0 (print)

      ISBN 978-0-7503-2217-1 (myPrint)

      ISBN 978-0-7503-2215-7 (mobi)

      DOI 10.1088/978-0-7503-2216-4

      Version: 20191201

      IOP ebooks

      British Library Cataloguing-in-Publication Data: A catalogue record for this book is available from the British Library.

      Published by IOP Publishing, wholly owned by The Institute of Physics, London

      IOP Publishing, Temple Circus, Temple Way, Bristol, BS1 6HG, UK

      US Office: IOP Publishing, Inc., 190 North Independence Mall West, Suite 601, Philadelphia, PA 19106, USA

      Contents

       Foreword

       Preface

       Acknowledgments

       Author biographies

       Introduction

       Part I Background

       1 Background knowledge

       1.1 Imaging principles and a priori information

       1.1.1 Overview

       1.1.2 Radon transform and non-ideality in data acquisition

       1.1.3 Bayesian reconstruction

       1.1.4 The human vision system

       1.1.5 Data decorrelation and whitening

       1.1.6 Sparse coding

       References

       2 Tomographic reconstruction based on a learned dictionary

       2.1 Prior information guided reconstruction

       2.2 Single-layer neural network

       2.2.1 Matching pursuit algorithm

       2.2.2 The K-SVD algorithm

       2.3 CT reconstruction via dictionary learning

       2.3.1 Statistic iterative reconstruction framework (SIR)

       2.3.2 Dictionary-based low-dose CT reconstruction

       2.4 Final remarks

       References

       3 Artificial neural networks

       3.1 Basic concepts

       3.1.1 Biological neural network

       3.1.2 Neuron models

       3.1.3 Activation function

       3.1.4 Discrete convolution and weights

       3.1.5 Pooling strategy

       3.1.6 Loss function

       3.1.7 Backpropagation algorithm

       3.1.8 Convolutional neural network

       3.2 Training, validation, and testing of an artificial neural network

       3.2.1 Training, validation, and testing datasets

       3.2.2 Training, validation, and testing processes

       3.2.3 Related concepts

       3.3 Typical artificial neural networks

       3.3.1 VGG network

       3.3.2