DOCSLIB.ORG

ABSTRACT LI, XINHAO. Development of Novel Machine Learning

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
ABSTRACT LI, XINHAO. Development of Novel Machine Learning ABSTRACT LI, XINHAO. Development of Novel Machine Learning Approaches and Exploration of Their Applications in Cheminformatics. (Under the direction of Dr. Christopher Gorman). Cheminformatics is the scientific field that develops and applies a combination of mathematics, informatics, machine learning, and other computational technologies to solve chemical problems. It aims at helping chemists in investigating and understanding complex chemical biological systems and guide the experimental design and decision making. With the rapid growing of chemical data (e.g., high-throughput screening, metabolomics, etc.), machine learning has become an important tool for exploring chemical space and mining chemical information. In this dissertation, we present five studies on developing novel machine learning approaches and exploring their applications in cheminformatics. One of the primary tasks of cheminformatics is to predict the physical, chemical, and biological properties of a given compound. Quantitative Structure Activity Relationship (QSAR) modeling relies on machine learning techniques to establish quantified links between molecular structures and their experimental properties/activities. In chapter 2, we developed a dual-layer hierarchical modeling method to fully integrate regression and classification QSAR models for assessing rat acute oral systemic toxicity, with respect to regulatory classifications of concern. The first layer of independent regression, binary and multiclass models (base models) were solely built using computed chemical descriptors/fingerprints. Then, a second layer of models (hierarchical models) were built by stacking all the cross-validated out-of-fold predictions from the base models. All models were validated using an external test set and we found that the hierarchical models did outperform the base models for all the three endpoints. The H-QSAR modeling method represents a promising approach for chemical toxicity prediction and more generally for stacking and blending individual QSAR models into more predictive ensemble models. In chapter 3, we proposed the Molecular Prediction Model Fine-Tuning (MolPMoFiT) approach, an effective transfer learning method based on self-supervised pre-training + task- specific fine-tuning for QSPR/QSAR modeling. It enables knowledge learned from large chemical data sets to transfer to smaller data sets, thereby improving the model performance and generalization. A large-scale molecular structure prediction model is pre-trained using one million unlabeled molecules from ChEMBL in a self-supervised learning manner and can then be fine- tuned on various QSPR/QSAR tasks for smaller chemical datasets with specific endpoints. The benchmark results show this transfer learning method can achieve strong performances compared to other state-of-the-art machine learning modeling techniques. SMILES-based deep learning models are slowly emerging as an important research topic in cheminformatics. In chapter 4, we introduced SMILES Pair Encoding (SPE), a data-driven tokenization algorithm. SPE learns a vocabulary of high frequency SMILES substrings from ChEMBL and then tokenizes new SMILES into a sequence of tokens for deep learning models. SPE augments the widely used atom-level tokenization by adding human-readable and chemically explainable SMILES substrings as tokens. Case studies show that SPE can achieve superior performances for both molecular generation (by boosting the validity and novelty of generated SMILES) as well as QSAR prediction tasks. In particular, we evaluated the performance of SPE- based QSAR prediction models using 24 benchmark datasets where SPE consistently either did match or outperform atom-level tokenization. SPE could represent a better tokenization method for the development of future deep learning applications in cheminformatics. Lead optimization endpoints usually contain a small set of congeneric molecules which limits the application of deep learning. Transfer learning enables training a deep learning model with less data. In chapter 5, we explored our transfer learning approach (see Chapter 3) for lead optimization endpoints. A set of lead-optimization-like benchmark datasets was created from 8 datasets that include pIC50 values against the protein targets. Two language models, a LSTM model and a RoBERTa model, were trained on 10M SMILES from ChEMBL and then used as the knowledge source for the downstream QSAR tasks. The results show transfer learning indeed provide performance gains for lead optimization endpoints. In chapter 6, we present CryptoChem, a new method and associated software to securely store and transfer information using chemicals. Relying on the concept of big chemical data, molecular descriptors and machine learning techniques, CryptoChem offers a highly complex and robust system with multiple layers of security for transmitting confidential information. The algorithm directly uses chemical structures and their properties as the central element of the secured storage. QSDR (Quantitative Structure-Data Relationship) models are used as private keys to encode and decode the data. The software is validated with a series of five datasets consisting of numerical and textual information with increasing size and complexity. © Copyright 2020 by Xinhao Li All Rights Reserved Development of Novel Machine Learning Approaches and Exploration of Their Applications in Cheminformatics by Xinhao Li A dissertation submitted to the Graduate Faculty of North Carolina State University in partial fulfillment of the requirements for the degree of Doctor of Philosophy Chemistry Raleigh, North Carolina 2020 APPROVED BY: _______________________________ _______________________________ Christopher Gorman Gavin Williams Committee Chair _______________________________ _______________________________ Caroline Proulx Brian Reich _______________________________ Denis Fourches External Member DEDICATION This dissertation is dedicated to Anqi Tu and my family. ii BIOGRAPHY Xinhao Li grew up in Dalian, China. With some very good experience in high school chemistry class, he decided to study chemistry in Beijing University of Chemical Technology where he got his bachelor’s and master’s degrees. During the master’s study, his research focus on organic synthesis. In 2017, he joined Dr. Denis Fourches’ lab at North Carolina State University and started to pursuit a Ph.D. with a focus on cheminformatics and machine learning. iii ACKNOWLEDGMENTS I would like to thank Dr. Denis Fourches for all the guidance, support, and encouragement he has given me for the past three years. His knowledge, vision and suggestions have been a valuable source of inspiration for my research. I would like to thank Dr. Christopher Gorman, Dr. Gavin Williams, Dr. Caroline Proulx and Dr. Brian Reich for being my committee members. I would like to thank Dr. Jamel Meslamani for his mentoring during my internship at GSK. I would also like to thank North Carolina State Chemistry Department and DARPA for the financial support. Finally, I would like to thank all my family and friends for their continued support over the years. I would like to especially thank Anqi Tu for the constant love and support along the way. iv TABLE OF CONTENTS LIST OF TABLES ..................................................................................................................... viii LIST OF FIGURES ..................................................................................................................... ix Chapter 1. Introduction ............................................................................................................... 1 1.1. Machine Learning ................................................................................................................ 2 1.2. QSAR Modeling ................................................................................................................... 3 References ................................................................................................................................... 8 Chapter 2. Hierarchical H-QSAR Modeling Approach for Integrating Binary/Multi Classification and Regression Models of Acute Oral Systemic Toxicity ............................... 10 Abstract ..................................................................................................................................... 11 2.1. Introduction ........................................................................................................................ 12 2.2. Materials and Methods ....................................................................................................... 15 2.3. Results ................................................................................................................................ 23 2.3.1. Chemical space of the full dataset ............................................................................... 23 2.3.2. Models development, validation, and comparison ...................................................... 24 2.3.3. Applicability Domain for Hierarchical Models ........................................................... 33 2.4. Discussion .......................................................................................................................... 37 2.4.1. Comparison with other methodologies recently reported in the literature .................. 39 2.5. Conclusion .........................................................................................................................
Load more

Recommended publications
  • Bottleneck Discovery and Overlay Management in Network Coded Peer-To-Peer Systems
    Bottleneck Discovery and Overlay Management in Network Coded Peer-to-Peer Systems ∗ Mahdi Jafarisiavoshani Christina Fragouli EPFL EPFL Switzerland Switzerland mahdi.jafari@epfl.ch christina.fragouli@epfl.ch Suhas Diggavi Christos Gkantsidis EPFL Microsoft Research Switzerland United Kingdom suhas.diggavi@epfl.ch [email protected] ABSTRACT 1. INTRODUCTION The performance of peer-to-peer (P2P) networks depends critically Peer-to-peer (P2P) networks have proved a very successful dis- on the good connectivity of the overlay topology. In this paper we tributed architecture for content distribution. The design philoso- study P2P networks for content distribution (such as Avalanche) phy of such systems is to delegate the distribution task to the par- that use randomized network coding techniques. The basic idea of ticipating nodes (peers) themselves, rather than concentrating it to such systems is that peers randomly combine and exchange linear a low number of servers with limited resources. Therefore, such a combinations of the source packets. A header appended to each P2P non-hierarchical approach is inherently scalable, since it ex- packet specifies the linear combination that the packet carries. In ploits the computing power and bandwidth of all the participants. this paper we show that the linear combinations a node receives Having addressed the problem of ensuring sufficient network re- from its neighbors reveal structural information about the network. sources, P2P systems still face the challenge of how to efficiently We propose algorithms to utilize this observation for topology man- utilize these resources while maintaining a decentralized operation. agement to avoid bottlenecks and clustering in network-coded P2P Central to this is the challenging management problem of connect- systems.
    [Show full text]
  • Developer Condemns City's Attitude Aican Appeal on Hold Avalanche
    Developer condemns city's attitude TERRACE -- Although city under way this spring, law when owners of lots adja- ment." said. council says it favours develop- The problem, he explained, cent to the sewer line and road However, alderman Danny ment, it's not willing to put its was sanitary sewer lines within developed their properties. Sheridan maintained, that is not Pointing out the development money where its mouth is, says the sub-division had to be hook- Council, however, refused the case. could still proceed if Shapitka a local developer. ed up to an existing city lines. both requests. The issue, he said, was paid the road and sewer connec- And that, adds Stan The nearest was :at Mountain • "It just seems the city isn't whether the city had subsidized tion costs, Sheridan said other Shapitka, has prompted him to Vista Drive, approximately too interested in lending any developers in the past -- "I'm developers had done so in the drop plans for what would have 850ft. fr0m-the:sou[hwest cur: type of assistance whatsoever," pretty sure it hasn't" -- and past. That included the city been the city's largest residential her of the development proper- Shapitka said, adding council whether it was going to do so in itself when it had developed sub-division project in many ty. appeared to want the estimated this case. "Council didn't seem properties it owned on the Birch years. Shapitka said he asked the ci- $500,000 increased tax base the willing to do that." Ave. bench and deJong Cres- In August of last year ty to build that line and to pave sub-division :would bring but While conceding Shapitka cent.
    [Show full text]
  • Survey of Verification and Validation Techniques for Small Satellite Software Development
    Survey of Verification and Validation Techniques for Small Satellite Software Development Stephen A. Jacklin NASA Ames Research Center Presented at the 2015 Space Tech Expo Conference May 19-21, Long Beach, CA Summary The purpose of this paper is to provide an overview of the current trends and practices in small-satellite software verification and validation. This document is not intended to promote a specific software assurance method. Rather, it seeks to present an unbiased survey of software assurance methods used to verify and validate small satellite software and to make mention of the benefits and value of each approach. These methods include simulation and testing, verification and validation with model-based design, formal methods, and fault-tolerant software design with run-time monitoring. Although the literature reveals that simulation and testing has by far the longest legacy, model-based design methods are proving to be useful for software verification and validation. Some work in formal methods, though not widely used for any satellites, may offer new ways to improve small satellite software verification and validation. These methods need to be further advanced to deal with the state explosion problem and to make them more usable by small-satellite software engineers to be regularly applied to software verification. Last, it is explained how run-time monitoring, combined with fault-tolerant software design methods, provides an important means to detect and correct software errors that escape the verification process or those errors that are produced after launch through the effects of ionizing radiation. Introduction While the space industry has developed very good methods for verifying and validating software for large communication satellites over the last 50 years, such methods are also very expensive and require large development budgets.
    [Show full text]
  • The Fourth Paradigm
    ABOUT THE FOURTH PARADIGM This book presents the first broad look at the rapidly emerging field of data- THE FOUR intensive science, with the goal of influencing the worldwide scientific and com- puting research communities and inspiring the next generation of scientists. Increasingly, scientific breakthroughs will be powered by advanced computing capabilities that help researchers manipulate and explore massive datasets. The speed at which any given scientific discipline advances will depend on how well its researchers collaborate with one another, and with technologists, in areas of eScience such as databases, workflow management, visualization, and cloud- computing technologies. This collection of essays expands on the vision of pio- T neering computer scientist Jim Gray for a new, fourth paradigm of discovery based H PARADIGM on data-intensive science and offers insights into how it can be fully realized. “The impact of Jim Gray’s thinking is continuing to get people to think in a new way about how data and software are redefining what it means to do science.” —Bill GaTES “I often tell people working in eScience that they aren’t in this field because they are visionaries or super-intelligent—it’s because they care about science The and they are alive now. It is about technology changing the world, and science taking advantage of it, to do more and do better.” —RhyS FRANCIS, AUSTRALIAN eRESEARCH INFRASTRUCTURE COUNCIL F OURTH “One of the greatest challenges for 21st-century science is how we respond to this new era of data-intensive
    [Show full text]
  • The Methbot Operation
    The Methbot Operation December 20, 2016 1 The Methbot Operation White Ops has exposed the largest and most profitable ad fraud operation to strike digital advertising to date. THE METHBOT OPERATION 2 Russian cybercriminals are siphoning At this point the Methbot operation has millions of advertising dollars per day become so embedded in the layers of away from U.S. media companies and the the advertising ecosystem, the only way biggest U.S. brand name advertisers in to shut it down is to make the details the single most profitable bot operation public to help affected parties take discovered to date. Dubbed “Methbot” action. Therefore, White Ops is releasing because of references to “meth” in its results from our research with that code, this operation produces massive objective in mind. volumes of fraudulent video advertising impressions by commandeering critical parts of Internet infrastructure and Information available for targeting the premium video advertising download space. Using an army of automated web • IP addresses known to belong to browsers run from fraudulently acquired Methbot for advertisers and their IP addresses, the Methbot operation agencies and platforms to block. is “watching” as many as 300 million This is the fastest way to shut down the video ads per day on falsified websites operation’s ability to monetize. designed to look like premium publisher inventory. More than 6,000 premium • Falsified domain list and full URL domains were targeted and spoofed, list to show the magnitude of impact enabling the operation to attract millions this operation had on the publishing in real advertising dollars.
    [Show full text]
  • The Failure of Mandated Disclosure
    BEN-SHAHAR_FINAL.DOC (DO NOT DELETE) 2/3/2011 11:53 AM ARTICLE THE FAILURE OF MANDATED DISCLOSURE † †† OMRI BEN-SHAHAR & CARL E. SCHNEIDER This Article explores the spectacular prevalence, and failure, of the single most common technique for protecting personal autonomy in modern society: mandated disclosure. The Article has four Parts: (1) a comprehensive sum- mary of the recurring use of mandated disclosures, in many forms and circums- tances, in the areas of consumer and borrower protection, patient informed con- sent, contract formation, and constitutional rights; (2) a survey of the empirical literature documenting the failure of the mandated disclosure regime in informing people and in improving their decisions; (3) an account of the multitude of reasons mandated disclosures fail, focusing on the political dy- namics underlying the enactments of these mandates, the incentives of disclosers to carry them out, and, most importantly, on the ability of disclosees to use them; and (4) an argument that mandated disclosure not only fails to achieve its stated goal but also leads to unintended consequences that often harm the very people it intends to serve. INTRODUCTION ......................................................................................649 A. The Argument ................................................................. 649 B. The Method ..................................................................... 651 C. The Style ......................................................................... 652 I. THE DISCLOSURE EMPIRE: THE PERVASIVENESS OF MANDATED DISCLOSURE ...................................................................................652 † Frank & Bernice J. Greenberg Professor of Law, University of Chicago. †† Chauncey Stillman Professor of Law & Professor of Internal Medicine, Universi- ty of Michigan. Helpful comments were provided by workshop participants at The University of Pennsylvania, Georgetown University, The University of Michigan, Tel- Aviv University, and the Federal Reserve Banks of Chicago and Cleveland.
    [Show full text]
  • An Avalanche Is Coming Higher Education and the Revolution Ahead
    AN AVALANCHE IS COMING HIGHER EDUCATION AND THE REVOLUTION AHEAD ESSAY Michael Barber Katelyn Donnelly Saad Rizvi Foreword by Lawrence Summers, President Emeritus, Harvard University March 2013 © IPPR 2013 Institute for Public Policy Research AN AVALANCHE IS COMING Higher education and the revolution ahead Michael Barber, Katelyn Donnelly, Saad Rizvi March 2013 ‘It’s tragic because, by my reading, should we fail to radically change our approach to education, the same cohort we’re attempting to “protect” could find that their entire future is scuttled by our timidity.’ David Puttnam Speech at Massachusetts Institute of Technology, June 2012 i ABOUT THE AUTHORS Sir Michael Barber is the chief education advisor at Pearson, leading Pearson’s worldwide programme of research into education policy and the impact of its products and services on learner outcomes. He chairs the Pearson Affordable Learning Fund, which aims to extend educational opportunity for the children of low-income families in the developing world. Michael also advises governments and development agencies on education strategy, effective governance and delivery. Prior to Pearson, he was head of McKinsey’s global education practice. He previously served the UK government as head of the Prime Minister’s Delivery Unit (2001–05) and as chief adviser to the secretary of state for education on school standards (1997–2001). Micheal is a visiting professor at the Higher School of Economics in Moscow and author of numerous books including Instruction to Deliver: Fighting to Improve Britain’s Public Services (2007) which was described by the Financial Times as ‘one of the best books about British government for many years’.
    [Show full text]
  • How Blockchain Will Disrupt Business
    SPECIAL REPORT How blockchain will disrupt business COPYRIGHT ©2018 CBS INTERACTIVE INC. ALL RIGHTS RESERVED. HOW BLOCKCHAIN WILL DISRUPT BUSINESS TABLE OF CONTENTS 03 Blockchain and business: Looking beyond the hype 18 Survey: Most professionals don’t have experience using blockchain, but many see its potential 19 Blockchain explained for non-engineers 22 Blockchain won’t save the world—but it might make it a better place 28 10 ways the enterprise could put blockchain to work 32 Why blockchain could transform the world financial system and what that would mean 36 How using blockchain in the supply chain could democratize innovation itself 40 How blockchain could change how we buy music, read news, and consume content 43 Could blockchain be the missing link in electronic voting? 2 COPYRIGHT ©2018 CBS INTERACTIVE INC. ALL RIGHTS RESERVED. HOW BLOCKCHAIN WILL DISRUPT BUSINESS BLOCKCHAIN AND BUSINESS: LOOKING BEYOND THE HYPE BY CHARLES MCLELLAN The term blockchain can elicit reactions ranging from a blank stare (from the majority of the general public) to evangelical fervour (from over-enthusiastic early adopters). But most people who know a bit about the technology detect a pungent whiff of hype, leavened with the suspicion that, when the dust settles, it may have a significant role to play as a component of digital transformation. The best-known example of blockchain technology in action is the leading cryptocurrency Bitcoin, but there are many more use cases—think of blockchain as the ‘operating system’ upon which different ‘applications’ (such as Bitcoin) can run. So, what is a blockchain? At heart, a blockchain is a special kind of database in which ‘blocks’ of sequential and immutable data pertaining to virtual or physical assets are linked via cryptographic hashes and distributed as an ever-growing ‘chain’ among multiple peer-to-peer ‘nodes’.
    [Show full text]
  • Mining and Ensembling Heterogeneous Information for Protein Interaction Predictions
    University of Wollongong Research Online Faculty of Engineering and Information Faculty of Engineering and Information Sciences - Papers: Part B Sciences 2020 HIME: Mining and Ensembling Heterogeneous Information for Protein Interaction Predictions Huaming Chen [email protected] Yaochu Jin [email protected] Lei Wang University of Wollongong, [email protected] Chi-Hung Chi [email protected] Jun Shen University of Wollongong, [email protected] Follow this and additional works at: https://ro.uow.edu.au/eispapers1 Part of the Engineering Commons, and the Science and Technology Studies Commons Recommended Citation Chen, Huaming; Jin, Yaochu; Wang, Lei; Chi, Chi-Hung; and Shen, Jun, "HIME: Mining and Ensembling Heterogeneous Information for Protein Interaction Predictions" (2020). Faculty of Engineering and Information Sciences - Papers: Part B. 4201. https://ro.uow.edu.au/eispapers1/4201 Research Online is the open access institutional repository for the University of Wollongong. For further information contact the UOW Library: [email protected] HIME: Mining and Ensembling Heterogeneous Information for Protein Interaction Predictions Abstract esearch on protein-protein interactions (PPIs) data paves the way towards understanding the mechanisms of infectious diseases, however improving the prediction performance of PPIs of inter- species remains a challenge. Since one single type of sequence data such as amino acid composition may be deficient for high-quality prediction of protein interactions, we have investigated a broader range of heterogeneous information of sequences data. This paper proposes a novel framework for PPIs prediction based on Heterogeneous Information Mining and Ensembling (HIME) process to effectively learn from the interaction data.
    [Show full text]
  • A Rich Seam: How New Pedagogies Find Deep
    A Rich Seam How New Pedagogies Find Deep Learning Authors Michael Fullan Maria Langworthy With the support of Foreword by Sir Michael Barber January 2014 A Rich Seam How New Pedagogies Find Deep Learning Michael Fullan & Maria Langworthy January 2014 A Rich Seam About The Authors About Pearson Michael Fullan Pearson is the world’s leading learning company. Our Michael Fullan, Order of Canada, is Professor Emeritus at education business combines 150 years of experience in the University of Toronto’s Ontario Institute for Studies publishing with the latest learning technology and online in Education. He has served as Special Adviser to Dalton support. We serve learners of all ages around the globe, McGuinty, the Premier of Ontario. He works around the employing 45,000 people in more than 70 countries, helping world to accomplish whole-system education reform in people to learn whatever, whenever and however they provinces, states and countries. He is a partner in the global GLSSWI;LIXLIVMX´WHIWMKRMRKUYEPM½GEXMSRWMRXLI9/ initiative New Pedagogies for Deep Learning. He is the supporting colleges in the US, training school leaders in the recipient of several honorary doctoral degrees. His award- Middle East or helping students in China learn English, we winning books have been published in many languages. His aim to help people make progress in their lives through most recent publications include Stratosphere: Integrating learning. technology, pedagogy, and change knowledge, Professional Capital of Teachers (with Andy Hargreaves), Motion Leadership in Action and The Principal: Three keys for maximizing impact. About Nesta See his website www.michaelfullan.ca 2IWXEMWXLI9/´WMRRSZEXMSRJSYRHEXMSR;ILIPTTISTPI Maria Langworthy and organisations bring great ideas to life.
    [Show full text]
  • Interactive Event-Driven Knowledge Discovery from Data Streams
    UC Irvine UC Irvine Electronic Theses and Dissertations Title Interactive Event-driven Knowledge Discovery from Data Streams Permalink https://escholarship.org/uc/item/8bc5k0j3 Author Jalali, Laleh Publication Date 2016 License https://creativecommons.org/licenses/by/4.0/ 4.0 Peer reviewed|Thesis/dissertation eScholarship.org Powered by the California Digital Library University of California UNIVERSITY OF CALIFORNIA, IRVINE Interactive Event-driven Knowledge Discovery from Data Streams DISSERTATION submitted in partial satisfaction of the requirements for the degree of DOCTOR OF PHILOSOPHY in Computer Science by Laleh Jalali Dissertation Committee: Professor Ramesh Jain, Chair Professor Gopi Meenakshisundaram Professor Nalini Venkatasubramanian 2016 © 2016 Laleh Jalali TABLE OF CONTENTS Page LIST OF FIGURES v LIST OF TABLES viii LIST OF ALGORITHMS ix ACKNOWLEDGMENTS x CURRICULUM VITAE xi ABSTRACT OF THE DISSERTATION xiv 1 Introduction 1 1.1 Data-driven vs. Hypothesis-driven . .3 1.2 Knowledge Discovery from Temporal Data . .5 1.3 Contributions . .7 1.4 Thesis Outline . .9 2 Understanding Knowledge Discovery Process 10 2.1 Knowledge Discovery Definitions . 11 2.2 New Paradigm: Actionable Knowledge Extraction . 13 2.3 Data Mining and Knowledge Discovery Software Tools . 15 2.4 Knowledge Discovery in Healthcare Applications . 20 2.5 Design Requirements for an Interactive Knowledge Discovery Framework . 23 3 Literature Review 26 3.1 Temporal Data Mining . 27 3.1.1 Definitions and Concepts . 27 3.1.2 Pattern Discovery . 29 3.1.3 Temporal Association Rules . 35 3.1.4 Time Series Data Mining . 36 3.1.5 Temporal Classification and Clustering . 37 3.2 Temporal Reasoning . 38 3.2.1 Interval-based Temporal Logic .
    [Show full text]
  • Time-Of-Flight Cameras and Microsoft Kinect TM
    C. Dal Mutto, P. Zanuttigh and G. M. Cortelazzo Time-of-Flight Cameras and Microsoft Kinect TM A user perspective on technology and applications January 24, 2013 Springer To my father Umberto, who has continuously stimulated my interest for research To Marco, who left us too early leaving many beautiful remembrances and above all the dawn of a new small life To my father Gino, professional sculptor, to whom I owe all my work about 3D Preface This book originates from the activity on 3D data processing and compression of still and dynamic scenes of the Multimedia Technology and Telecommunications Laboratory (LTTM) of the Department of Information Engineering of the Univer- sity of Padova, which brought us to use Time-of-Flight cameras since 2008 and lately the Microsoft KinectTM depth measurement instrument. 3D data acquisition and visualization are topics of computer vision and computer graphics interest with several aspects relevant also for the telecommunication community active on multi- media. This book treats a number of methodological questions essentially concerning the best usage of the 3D data made available by ToF cameras and Kinect TM , within a general approach valid independently of the specific depth measurement instrument employed. The practical exemplification of the results with actual data given in this book does not only confirm the effectiveness of the presented methods, but it also clarifies them and gives the reader a sense for their concrete possibilities. The reader will note that the ToF camera data presented in this book are obtained by the Mesa SR4000. This is because Mesa Imaging kindly made available their ToF camera and their guidance for its best usage in the experiments of this book.
    [Show full text]