The Prac)cal Applica)on of Knowledge Discovery to Image Data: A Prac))oners View in The Context of Medical Image Mining

Size: px
Start display at page:

Download "The Prac)cal Applica)on of Knowledge Discovery to Image Data: A Prac))oners View in The Context of Medical Image Mining"

Transcription

1 The Prac)cal Applica)on of Knowledge Discovery to Image Data: A Prac))oners View in The Context of Medical Image Mining Frans Coenen ( 10th Interna+onal Conference on Natural Computa+on (ICNC 2014) 11th Interna+onal Conference on Fuzzy Systems and Knowledge Discovery (FSKD 2014)

2 The Challenge of Knowledge Discovery We have many well established techniques for conduc+ng Knowledge Discovery (KD). The issue is the end- to- end KD process. More specifically the prepara+on of data so that KD algorithms can be applied. This is rela+vely straight forward given (tradi+onal) tabular data. This is much more challenging in the context of complex data such as: graphs, free text, video and image data.

3 Presenta)on Focus Image data, both 2D and 3D. Specifically medical image data. Classifica+on (Predic+on) of condi+ons. Classifica+on genera+on algorithms tend to require feature vector (signature) represented training data (but not exclusively so). Three genres of representa+on considered: (i) Sta+s+cal, (ii) Point Series and (iii) Graph Based.

4 Mo)va)on Clinical decisions are regularly made with the support of medical imagery of some sort (MRI, OCT, etc.). Our ability to collect medical imagery has enhanced rapidly over the last decade (we can do it cheaper and faster). We have seen rapid growth in the global medical imaging sensor market Analysis is s+ll mostly conducted manually, there is very lixle soyware automa+on (although some suppor+ng tools, e.g. Brain Voyager).

5 Classifica)on Process Labelled Training Data Data processing Classifier Generator Unlabelled Sample Data processing Classifier Labelled Sample

6 Example Applica)ons One (2D Re)na Image Data) (b) (a) Example of a 2- D re0na images: (a) Normal re0na, (b) re0nas featuring AMD. We are interested in detec+ng condi+ons such as Age related Macular Degenera+on (AMD) and Diabe+c Re+nopathy (DR).

7 Example Applica)on Two (2D and 3D MRI Brain Scan Data) (a) (b) (c) Example 3- D brain MRI scan: (a) SagiEal plane, (b) Coronal plane, (c) Transverse plane (each scan consists of 256 slices, 768 slices in total). Interested in detec+ng condi+ons such as epilepsy and handedness according to par+cular objects within MRI images (ventricles, corpus callosum).

8 Example Applica)on Three (OCT Re)na Volumes) (a) (b)! Example of a 3- D OCT re0na volumes: (a) Normal re0na, (b) re0na featuring AMD. About 19 slices per volume ( pixels). Again interested in detec+ng condi+ons such as AMD and DR.!

9 Image Capture (a) Re0na Fundus Camera. (b) Magne0c Resonance Imagery (MRI) Machine. (c) Op0cal Coherence Tomography (OCT) set up.

10 Image Representa+on for KD Image Rep. for KD Global (Whole Image based) Sta+s+cal Local (Region/Object based) Graph based Sta+s+cal Set of regions objects Individual (regions/ objects) Point series/clouds Graph based Sta+s+cal

11 Image Representa+on for KD cont. Two approaches: (i) Global (whole image based) and (ii) Local (region or object based). The laxer may require segmenta+on which is an en+re research issue in its own right (possibly outside of the scope of KD). Three (broad) techniques: 1. Sta+s+cal. 2. Point Series. 3. Graph based.

12 Image Representa+on for KD cont. Note that our representa+on needs to be understandable to knowledge discovery algorithms, it does not necessarily have to be understandable to human observers. Although explana+on genera+on is an issue.

13 Image Representa+on for KD Techniques 1. Sta+s+cal. 2. Point Series. 3. Graph based.

14 Technique One (Sta)s)cal Techniques) Simplest approach and easy to represent in terms of a feature space directly compa+ble with classifier genera+on/applica+on. First Order Sta+s+cal func+ons such as the mean, variance and standard devia+on of the intensity or RGB colour values. Alterna+vely moments about some point or axis. Second order sta+s+cal func+ons applied to an intermediate representa+on (co- occurrence matrices, gradient analysis, Hough transforms). General applicability.

15 Technique Two (Point Series) Many second order sta+s+cal techniques lend themselves to representa+on in the form of histograms For example histogram of intensity values, LPBs or orienta+on gradients. Histograms can of course be directly translated into a feature vector representa+on. Alterna+vely, they can be viewed as point series or point curves.

16 Technique Two (Point Series) Given a collec+on of class labelled point series, represen+ng a set of images, a new image can be classified directly using (say) the well established KNN algorithm (a) Histogram of 20 bins. (b) Resul+ng point series.

17 Classifica)on Using Point Series cont. When using algorithms such as KNN we need an appropriate similarity measure (in the case of our point series Euclidean distance does not do the trick). We can look to work on point series comparison techniques, for example the Dynamic Time Warping (DTW) technique which produces a warping path distance defining the difference between two series. DTW has the added advantage that the point series to be compared do not have to be of the same length.

18 Genera+on of Point Series The histogram based approaches to point series genera+on is will suited to the sta+s+cal based techniques (whole image or region based). It is not so well suited to object based representa+on. In this case some alterna+ve mechanism is required to generate the desired point series

19 Example Point Series Genera)on (a) Right ventricle from a 3D MRI Brain Scan Volume. (b) Disc based point series genera0on. (c) Spoke based point series genera0on (conducted in three planes).

20 Second Example Using The Corpus Callosum Alterna+ve 2D example using a segmented object (the corpus callosum) extracted from a 2D MRI brain scan image. (a) 2D Radius intersec0on based point series genera0on. (b) Resul0ng point series.

21 Technique Three (Graph Based) A popular method for represen+ng images is to apply some form of hierarchical decomposi+on and to store the result in a quad- tree (2D image data) or oct- tree (3D image data). Issues with: 1. boundary problem where the object of interest is in different branches of the tree. 2. When to stop the decomposi+on (cri+cal func+on to measure homogeneity or a pre- specified maximum level of decomposi+on).

22 Example Decomposi+on One (a) Corpus callosum in a 2D MRI brain scan. (b) Segmented Corpus callosum. (c) Decomposed Corpus Callosum (level = 3). (d) Quad tree representa0on of the Corpus Callosum.

23 Example Decomposi+on Two (a) Whole image decomposi0on using an alterna0ve approach (level=4). (b) Resul0ng tree representa0on.

24 Example Decomposi+on three (a) Extra ``box in middle. (b) Combined oct- tree and quad tree decomposi0on. Used with respect to OCT re+na volumes where one dimension is significantly longer than the other two. Also an axempt to get round the boundary problem. An alterna+ve approach to the boundary problem would be to use some sort of fuzzy rela+onship func+on.

25 Genera)ng a Feature Space from a Collec)on of (Transac)on) Graphs Apply a Frequent Sub- graph Mining (FSM) algorithm to the data. Frequent defined by some threshold σ. Popular FSM algorithm is gspan. Each frequently occurring sub graph is then a dimension in a feature space. This feature space can then be used define feature vectors (binary or real valued) for the ini+al image set, which can be input to any number of classifier generators (Decision Tree generators, SVMs, Naïve Bayes, Classifica+on Associa+on Rule Miners, Etc.).

26 Classifica)on Process (with frequent sub graph graph mining) Labelled Training Data Data processing (graph genera+on) Frequent Sub- graph mining Feature Space Classifier Generator Unlabelled Sample Data processing (graph genera+on) Classifier Labelled Sample

27 Evalua+on (2D Re+na Images) Sensi)vity Specificity Accuracy AUC FNR Sta)s)cal Point Series Graph Based Whole image classifica+on. Two classes (AMD v. not AMD). 394 images, 165 that featured AMD, and 229 that did not. Re0na image featuring Age related Macular Degenera0on (AMD)

28 Evalua+on (2D Re+na Images cont.) Sensi)vity AMD Sensi)vity Other Whole image classifica+on. Three classes (AMD, DR and normal). 394 images, 165 that featured AMD, and 131 that featured DR and 98 normal. Specif- icity Accu- racy AUC FNR Sta)s)cal Point Series Graph Based Re0na image featuring Age related Macular Degenera0on (AMD) Re0na image featuring Diabe0c Re0nopathy (DR)

29 Evalua+on (2D Musicians MRI Images) Sta)s)cal (Hough Transform) Sta)s)cal (Zernicke Moments) Accuracy Sensi)vity Specificity Point Series Graph Based MRI scans, 53 represen+ng musicians and 53 non- musicians (i.e. two equal classes). Study of interest because of the conjecture that size and shape of the corpus callosum reflects human characteris+cs such as musical or mathema+cal ability.

30 Evalua+on (2D Handedness MRI Images) Sta)s)cal (Hough Transform) Sta)s)cal (Zernicke Moments) Accuracy Sensi)vity Specificity Point Series Graph Based MRI brain scans, 42 right handed and 40 ley handed. Study of interest because of the conjecture that size and shape of the corpus callosum reflects human characteris+cs such as handedness.

31 Evalua+on (2D Epilepsy MRI Images) Sta)s)cal (Hough Transform) Sta)s)cal (Zernicke Moments) Accuracy Sensi)vity Specificity Point Series Graph Based MRI brain scans, 106 from the musicians study plus 106 epilepsy scans. The objec+ve was to seek support for the conjecture that the shape and size of the corpus callosum is influence by condi+ons such as epilepsy.

32 Evalua+on (3D Epilepsy MRI Images) 210 MRI brain scans, 106 from musicians study, plus 106 epilepsy scans. Used the ley and right ventricles as objet of interest. Accuracy Sensi)vity Specificity Point Series (Disc Based) Point Series (Spoke Based 2 Spacing) Point Series (Spoke Based 3 Spacing) Graph Based (σ=30%) Spoke Based Point Series Genera+on (in three planes).

33 Evalua+on (3D OCT Re+na Volumes) Accuracy Sensi)vity Specificity AUC Sta)s)cal (oct- tree) Sta)s)cal (oct- tree with overlap) Graph Based (oct- tree with overlap) D OCT volumes, 2 classes, 68 normal and 52 AMD. Level one 9- tree

34 Conclusions Number of alterna+ve (whole and region/object based) image representa+ons considered, directed at capturing salient features in 2D and 3D image data for classifica+on/predic+on purposes. Three categories of representa+on considered: (i) sta+s+cal, (ii) point series and (iii) tree based. Evaluated with respect to a number of medical applica+on domains. No best representa+on L.

35 Further Work S+ll issues with explana+on genera+on. Lot of scope for alterna+ve representa+ons, especially fuzzy approaches. Lots of scope for further applica+on. (a) Mammogram (a) Hand X- ray

36 Acknowledgements Abdulrahman Albarrak [1], Ashraf Elsayed [2], Marta García- Fiñana [3], Hanafi Hijazi [4], Vanessa Sluming [5], Akadej Udomchaiporn [1] and Yalin Zheng [6]. [1] Dept. of Computer Science, University of Liverpool, Liverpool, UK. [2] Dept. of Computer Science, University of Alexandria, Egypt. [3] Dept. of Biosta+s+cs, University of Liverpool, Liverpool, UK. [4] School of Engineering and Informa+on Technology, University of Malaysia Sabah, Malaysia. [5] School of Health Science, University of Liverpool, Liverpool, UK. [6] Dept. of Eye and Vision Science, Royal Liverpool University Hospital, UK.

The Prac)cal Applica)on of Knowledge Discovery to Image Data: A Prac))oners View in the Context of Medical Image Diagnos)cs

The Prac)cal Applica)on of Knowledge Discovery to Image Data: A Prac))oners View in the Context of Medical Image Diagnos)cs The Prac)cal Applica)on of Knowledge Discovery to Image Data: A Prac))oners View in the Context of Medical Image Diagnos)cs Frans Coenen (http://cgi.csc.liv.ac.uk/~frans/) University of Mauri0us, June

More information

3-D MRI Brain Scan Classification Using A Point Series Based Representation

3-D MRI Brain Scan Classification Using A Point Series Based Representation 3-D MRI Brain Scan Classification Using A Point Series Based Representation Akadej Udomchaiporn 1, Frans Coenen 1, Marta García-Fiñana 2, and Vanessa Sluming 3 1 Department of Computer Science, University

More information

3-D MRI Brain Scan Feature Classification Using an Oct-tree Representation

3-D MRI Brain Scan Feature Classification Using an Oct-tree Representation 3-D MRI Brain Scan Feature Classification Using an Oct-tree Representation Akadej Udomchaiporn 1, Frans Coenen 1, Marta García-Fiñana 2, and Vanessa Sluming 3 1 Department of Computer Science, University

More information

3-D Volume of Interest Based Image Classification

3-D Volume of Interest Based Image Classification 3-D Volume of Interest Based Image Classification Akadej Udomchaiporn 1, Frans Coenen 1, Marta García-Fiñana 2, and Vanessa Sluming 3 1 Department of Computer Science, University of Liverpool, Liverpool,

More information

Region Of Interest Based Image Classification: A Study in MRI Brain Scan Categorization

Region Of Interest Based Image Classification: A Study in MRI Brain Scan Categorization Region Of Interest Based Image Classification: A Study in MRI Brain Scan Categorization Thesis submitted in accordance with the requirements of the University of Liverpool for the degree of Doctor in Philosophy

More information

Alignment and Image Comparison

Alignment and Image Comparison Alignment and Image Comparison Erik Learned- Miller University of Massachuse>s, Amherst Alignment and Image Comparison Erik Learned- Miller University of Massachuse>s, Amherst Alignment and Image Comparison

More information

Region Of Interest Based Image Classification Using Time Series Analysis

Region Of Interest Based Image Classification Using Time Series Analysis Region Of Interest Based Image Classification Using Time Series Analysis A. Elsayed, F. Coenen, M. García-Fiñana and V. Sluming Abstract An approach to Region Of Interest Based Image Classification (ROIBIC),

More information

Alignment and Image Comparison. Erik Learned- Miller University of Massachuse>s, Amherst

Alignment and Image Comparison. Erik Learned- Miller University of Massachuse>s, Amherst Alignment and Image Comparison Erik Learned- Miller University of Massachuse>s, Amherst Alignment and Image Comparison Erik Learned- Miller University of Massachuse>s, Amherst Alignment and Image Comparison

More information

Search Engines. Informa1on Retrieval in Prac1ce. Annota1ons by Michael L. Nelson

Search Engines. Informa1on Retrieval in Prac1ce. Annota1ons by Michael L. Nelson Search Engines Informa1on Retrieval in Prac1ce Annota1ons by Michael L. Nelson All slides Addison Wesley, 2008 Evalua1on Evalua1on is key to building effec$ve and efficient search engines measurement usually

More information

Cross- Valida+on & ROC curve. Anna Helena Reali Costa PCS 5024

Cross- Valida+on & ROC curve. Anna Helena Reali Costa PCS 5024 Cross- Valida+on & ROC curve Anna Helena Reali Costa PCS 5024 Resampling Methods Involve repeatedly drawing samples from a training set and refibng a model on each sample. Used in model assessment (evalua+ng

More information

CAP5415-Computer Vision Lecture 13-Support Vector Machines for Computer Vision Applica=ons

CAP5415-Computer Vision Lecture 13-Support Vector Machines for Computer Vision Applica=ons CAP5415-Computer Vision Lecture 13-Support Vector Machines for Computer Vision Applica=ons Guest Lecturer: Dr. Boqing Gong Dr. Ulas Bagci bagci@ucf.edu 1 October 14 Reminders Choose your mini-projects

More information

Volume Visualiza0on. Today s Class. Grades & Homework feedback on Homework Submission Server

Volume Visualiza0on. Today s Class. Grades & Homework feedback on Homework Submission Server 11/3/14 Volume Visualiza0on h3p://imgur.com/trjonqk h3p://i.imgur.com/zcjc9kp.jpg Today s Class Grades & Homework feedback on Homework Submission Server Everything except HW4 (didn t get to that yet) &

More information

Visualizing Logical Dependencies in SWRL Rule Bases

Visualizing Logical Dependencies in SWRL Rule Bases Visualizing Logical Dependencies in SWRL Rule Bases Saeed Hassanpour, Mar:n J. O Connor and Amar K. Das Stanford Center for Biomedical Informa:cs Research MSOB X215, 251 Campus Drive, Stanford, California,

More information

Introduc)on to Informa)on Visualiza)on

Introduc)on to Informa)on Visualiza)on Introduc)on to Informa)on Visualiza)on Seeing the Science with Visualiza)on Raw Data 01001101011001 11001010010101 00101010100110 11101101011011 00110010111010 Visualiza(on Applica(on Visualiza)on on

More information

Image Classification for Age-related Macular Degeneration Screening using Hierarchical Image Decompositions and Graph Mining

Image Classification for Age-related Macular Degeneration Screening using Hierarchical Image Decompositions and Graph Mining Image Classification for Age-related Macular Degeneration Screening using Hierarchical Image Decompositions and Graph Mining Mohd Hanafi Ahmad Hijazi 1,3, Chuntao Jiang 1, Frans Coenen 1, and Yalin Zheng

More information

STA 4273H: Sta-s-cal Machine Learning

STA 4273H: Sta-s-cal Machine Learning STA 4273H: Sta-s-cal Machine Learning Russ Salakhutdinov Department of Statistics! rsalakhu@utstat.toronto.edu! h0p://www.cs.toronto.edu/~rsalakhu/ Lecture 3 Parametric Distribu>ons We want model the probability

More information

Search Engines. Informa1on Retrieval in Prac1ce. Annotations by Michael L. Nelson

Search Engines. Informa1on Retrieval in Prac1ce. Annotations by Michael L. Nelson Search Engines Informa1on Retrieval in Prac1ce Annotations by Michael L. Nelson All slides Addison Wesley, 2008 Retrieval Models Provide a mathema1cal framework for defining the search process includes

More information

COMP 9517 Computer Vision

COMP 9517 Computer Vision COMP 9517 Computer Vision Pa6ern Recogni:on (1) 1 Introduc:on Pa#ern recogni,on is the scien:fic discipline whose goal is the classifica:on of objects into a number of categories or classes Pa6ern recogni:on

More information

Global Journal of Engineering Science and Research Management

Global Journal of Engineering Science and Research Management ADVANCED K-MEANS ALGORITHM FOR BRAIN TUMOR DETECTION USING NAIVE BAYES CLASSIFIER Veena Bai K*, Dr. Niharika Kumar * MTech CSE, Department of Computer Science and Engineering, B.N.M. Institute of Technology,

More information

ICD- 10 Presenta/on #1 for Revolu/onEHR Users June 3, 2015

ICD- 10 Presenta/on #1 for Revolu/onEHR Users June 3, 2015 ICD- 10 Presenta/on #1 for Revolu/onEHR Users June 3, 2015 ScoA Jens, OD, FAAO CEO, Revolu/onEHR Discussion Points Overview of ICD Anatomy of ICD- 10 Choosing ICD- 10 Codes Summarizing Steps You Will Experience

More information

Age-related Macular Degeneration Identification In Volumetric Optical Coherence Tomography Using Decomposition and Local Feature Extraction

Age-related Macular Degeneration Identification In Volumetric Optical Coherence Tomography Using Decomposition and Local Feature Extraction ALBARRAK, COENEN, ZHENG: AMD IDENTIFICATION IN 3D OCT 1 Age-related Macular Degeneration Identification In Volumetric Optical Coherence Tomography Using Decomposition and Local Feature Extraction Abdulrahman

More information

Pa#ern Recogni-on for Neuroimaging Toolbox

Pa#ern Recogni-on for Neuroimaging Toolbox Pa#ern Recogni-on for Neuroimaging Toolbox Pa#ern Recogni-on Methods: Basics João M. Monteiro Based on slides from Jessica Schrouff and Janaina Mourão-Miranda PRoNTo course UCL, London, UK 2017 Outline

More information

An Latent Feature Model for

An Latent Feature Model for An Addi@ve Latent Feature Model for Mario Fritz UC Berkeley Michael Black Brown University Gary Bradski Willow Garage Sergey Karayev UC Berkeley Trevor Darrell UC Berkeley Mo@va@on Transparent objects

More information

Quick review: Data Mining Tasks... Classifica(on [Predic(ve] Regression [Predic(ve] Clustering [Descrip(ve] Associa(on Rule Discovery [Descrip(ve]

Quick review: Data Mining Tasks... Classifica(on [Predic(ve] Regression [Predic(ve] Clustering [Descrip(ve] Associa(on Rule Discovery [Descrip(ve] Evaluation Quick review: Data Mining Tasks... Classifica(on [Predic(ve] Regression [Predic(ve] Clustering [Descrip(ve] Associa(on Rule Discovery [Descrip(ve] Classification: Definition Given a collec(on

More information

CS395T Visual Recogni5on and Search. Gautam S. Muralidhar

CS395T Visual Recogni5on and Search. Gautam S. Muralidhar CS395T Visual Recogni5on and Search Gautam S. Muralidhar Today s Theme Unsupervised discovery of images Main mo5va5on behind unsupervised discovery is that supervision is expensive Common tasks include

More information

CS6200 Informa.on Retrieval. David Smith College of Computer and Informa.on Science Northeastern University

CS6200 Informa.on Retrieval. David Smith College of Computer and Informa.on Science Northeastern University CS6200 Informa.on Retrieval David Smith College of Computer and Informa.on Science Northeastern University Query Process Retrieval Models Provide a mathema.cal framework for defining the search process

More information

Accelerating Satellite Image Based Large- Scale Settlement Detection with GPU!

Accelerating Satellite Image Based Large- Scale Settlement Detection with GPU! Accelerating Satellite Image Based Large- Scale Settlement Detection with GPU! Dilip%R.%Patlolla% Anil%M.%Cheriyadat% Eddie%A.%Bright% Jeane9e%E.%Weaver% % Oak%Ridge%Na?onal%Laboratory% Oak%Ridge,%TN%%

More information

MapReduce, Apache Hadoop

MapReduce, Apache Hadoop Czech Technical University in Prague, Faculty of Informaon Technology MIE-PDB: Advanced Database Systems hp://www.ksi.mff.cuni.cz/~svoboda/courses/2016-2-mie-pdb/ Lecture 12 MapReduce, Apache Hadoop Marn

More information

Tumor Detection and classification of Medical MRI UsingAdvance ROIPropANN Algorithm

Tumor Detection and classification of Medical MRI UsingAdvance ROIPropANN Algorithm International Journal of Engineering Research and Advanced Technology (IJERAT) DOI:http://dx.doi.org/10.31695/IJERAT.2018.3273 E-ISSN : 2454-6135 Volume.4, Issue 6 June -2018 Tumor Detection and classification

More information

MapReduce, Apache Hadoop

MapReduce, Apache Hadoop NDBI040: Big Data Management and NoSQL Databases hp://www.ksi.mff.cuni.cz/ svoboda/courses/2016-1-ndbi040/ Lecture 2 MapReduce, Apache Hadoop Marn Svoboda svoboda@ksi.mff.cuni.cz 11. 10. 2016 Charles University

More information

BoW model. Textual data: Bag of Words model

BoW model. Textual data: Bag of Words model BoW model Textual data: Bag of Words model With text, categoriza9on is the task of assigning a document to one or more categories based on its content. It is appropriate for: Detec9ng and indexing similar

More information

CS 6140: Machine Learning Spring 2017

CS 6140: Machine Learning Spring 2017 CS 6140: Machine Learning Spring 2017 Instructor: Lu Wang College of Computer and Informa@on Science Northeastern University Webpage: www.ccs.neu.edu/home/luwang Email: luwang@ccs.neu.edu Logis@cs Grades

More information

CHAPTER-1 INTRODUCTION

CHAPTER-1 INTRODUCTION CHAPTER-1 INTRODUCTION 1.1 Fuzzy concept, digital image processing and application in medicine With the advancement of digital computers, it has become easy to store large amount of data and carry out

More information

Ontology engineering. Valen.na Tamma. Based on slides by A. Gomez Perez, N. Noy, D. McGuinness, E. Kendal, A. Rector and O. Corcho

Ontology engineering. Valen.na Tamma. Based on slides by A. Gomez Perez, N. Noy, D. McGuinness, E. Kendal, A. Rector and O. Corcho Ontology engineering Valen.na Tamma Based on slides by A. Gomez Perez, N. Noy, D. McGuinness, E. Kendal, A. Rector and O. Corcho Summary Background on ontology; Ontology and ontological commitment; Logic

More information

Decision Trees: Representa:on

Decision Trees: Representa:on Decision Trees: Representa:on Machine Learning Fall 2017 Some slides from Tom Mitchell, Dan Roth and others 1 Key issues in machine learning Modeling How to formulate your problem as a machine learning

More information

Informa(on Retrieval

Informa(on Retrieval Introduc*on to Informa(on Retrieval CS276: Informa*on Retrieval and Web Search Pandu Nayak and Prabhakar Raghavan Lecture 12: Clustering Today s Topic: Clustering Document clustering Mo*va*ons Document

More information

Rapid Extraction and Updating Road Network from LIDAR Data

Rapid Extraction and Updating Road Network from LIDAR Data Rapid Extraction and Updating Road Network from LIDAR Data Jiaping Zhao, Suya You, Jing Huang Computer Science Department University of Southern California October, 2011 Research Objec+ve Road extrac+on

More information

Association Rule Mining in The Wider Context of Text, Images and Graphs

Association Rule Mining in The Wider Context of Text, Images and Graphs Association Rule Mining in The Wider Context of Text, Images and Graphs Frans Coenen Department of Computer Science UKKDD 07, April 2007 PRESENTATION OVERVIEW Motivation. Association Rule Mining (quick

More information

Manipula0on Algorithms Mo0on Planning. Mo#on Planning I. Katharina Muelling (NREC, Carnegie Mellon University) 1

Manipula0on Algorithms Mo0on Planning. Mo#on Planning I. Katharina Muelling (NREC, Carnegie Mellon University) 1 16-843 Manipula0on Algorithms Mo0on Planning Mo#on Planning I Katharina Muelling (NREC, Carnegie Mellon University) 1 Configura0on Space Obstacles Star Algorithm Convex robot, transla#on C obs : convex

More information

Introduc)on to Probabilis)c Latent Seman)c Analysis. NYP Predic)ve Analy)cs Meetup June 10, 2010

Introduc)on to Probabilis)c Latent Seman)c Analysis. NYP Predic)ve Analy)cs Meetup June 10, 2010 Introduc)on to Probabilis)c Latent Seman)c Analysis NYP Predic)ve Analy)cs Meetup June 10, 2010 PLSA A type of latent variable model with observed count data and nominal latent variable(s). Despite the

More information

Informa(on Retrieval

Informa(on Retrieval Introduc*on to Informa(on Retrieval Clustering Chris Manning, Pandu Nayak, and Prabhakar Raghavan Today s Topic: Clustering Document clustering Mo*va*ons Document representa*ons Success criteria Clustering

More information

CITS4009 Introduc0on to Data Science

CITS4009 Introduc0on to Data Science School of Computer Science and Software Engineering CITS4009 Introduc0on to Data Science SEMESTER 2, 2017: CHAPTER 3 EXPLORING DATA 1 Chapter Objec0ves Using summary sta.s.cs to explore data Exploring

More information

What were his cri+cisms? Classical Methodologies:

What were his cri+cisms? Classical Methodologies: 1 2 Classifica+on In this scheme there are several methodologies, such as Process- oriented, Blended, Object Oriented, Rapid development, People oriented and Organisa+onal oriented. According to David

More information

Machine Learning Crash Course: Part I

Machine Learning Crash Course: Part I Machine Learning Crash Course: Part I Ariel Kleiner August 21, 2012 Machine learning exists at the intersec

More information

Spa$al Analysis and Modeling (GIST 4302/5302) Guofeng Cao Department of Geosciences Texas Tech University

Spa$al Analysis and Modeling (GIST 4302/5302) Guofeng Cao Department of Geosciences Texas Tech University Spa$al Analysis and Modeling (GIST 4302/5302) Guofeng Cao Department of Geosciences Texas Tech University Class Outlines Spatial Point Pattern Regional Data (Areal Data) Continuous Spatial Data (Geostatistical

More information

2. LITERATURE REVIEW

2. LITERATURE REVIEW 2. LITERATURE REVIEW CBIR has come long way before 1990 and very little papers have been published at that time, however the number of papers published since 1997 is increasing. There are many CBIR algorithms

More information

Computer Programming-I. Developed by: Strawberry

Computer Programming-I. Developed by: Strawberry Computer Programming-I Objec=ve of CP-I The course will enable the students to understand the basic concepts of structured programming. What is programming? Wri=ng a set of instruc=ons that computer use

More information

Object Oriented Design (OOD): The Concept

Object Oriented Design (OOD): The Concept Object Oriented Design (OOD): The Concept Objec,ves To explain how a so8ware design may be represented as a set of interac;ng objects that manage their own state and opera;ons 1 Topics covered Object Oriented

More information

Available Online through

Available Online through Available Online through www.ijptonline.com ISSN: 0975-766X CODEN: IJPTFI Research Article ANALYSIS OF CT LIVER IMAGES FOR TUMOUR DIAGNOSIS BASED ON CLUSTERING TECHNIQUE AND TEXTURE FEATURES M.Krithika

More information

Organized Segmenta.on

Organized Segmenta.on Organized Segmenta.on Alex Trevor, Georgia Ins.tute of Technology PCL TUTORIAL @ICRA 13 Overview Mo.va.on Connected Component Algorithm Planar Segmenta.on & Refinement Euclidean Clustering Timing Results

More information

Lecture 6: Multimedia Information Retrieval Dr. Jian Zhang

Lecture 6: Multimedia Information Retrieval Dr. Jian Zhang Lecture 6: Multimedia Information Retrieval Dr. Jian Zhang NICTA & CSE UNSW COMP9314 Advanced Database S1 2007 jzhang@cse.unsw.edu.au Reference Papers and Resources Papers: Colour spaces-perceptual, historical

More information

Histograms. h(r k ) = n k. p(r k )= n k /NM. Histogram: number of times intensity level rk appears in the image

Histograms. h(r k ) = n k. p(r k )= n k /NM. Histogram: number of times intensity level rk appears in the image Histograms h(r k ) = n k Histogram: number of times intensity level rk appears in the image p(r k )= n k /NM normalized histogram also a probability of occurence 1 Histogram of Image Intensities Create

More information

ML4Bio Lecture #1: Introduc3on. February 24 th, 2016 Quaid Morris

ML4Bio Lecture #1: Introduc3on. February 24 th, 2016 Quaid Morris ML4Bio Lecture #1: Introduc3on February 24 th, 216 Quaid Morris Course goals Prac3cal introduc3on to ML Having a basic grounding in the terminology and important concepts in ML; to permit self- study,

More information

COMP Associa0on Rules

COMP Associa0on Rules COMP 4601 Associa0on Rules 1 Road map Basic concepts Apriori algorithm Different data formats for mining Mining with mul0ple minimum supports Mining class associa0on rules Summary 2 What Is Frequent Pattern

More information

WP7: Patents Case Study

WP7: Patents Case Study MOLTO WP7: Patents Case Study Meritxell Gonzàlez Bermúdez 2nd Year Review Barcelona, March 20th, 2012 Objectives To create a prototype of MT and NL retrieval of patents in the bio- medical & pharmaceu;cal

More information

Logis&c Regression. Aar$ Singh & Barnabas Poczos. Machine Learning / Jan 28, 2014

Logis&c Regression. Aar$ Singh & Barnabas Poczos. Machine Learning / Jan 28, 2014 Logis&c Regression Aar$ Singh & Barnabas Poczos Machine Learning 10-701/15-781 Jan 28, 2014 Linear Regression & Linear Classifica&on Weight Height Linear fit Linear decision boundary 2 Naïve Bayes Recap

More information

CHAPTER-4 LOCALIZATION AND CONTOUR DETECTION OF OPTIC DISK

CHAPTER-4 LOCALIZATION AND CONTOUR DETECTION OF OPTIC DISK CHAPTER-4 LOCALIZATION AND CONTOUR DETECTION OF OPTIC DISK Ocular fundus images can provide information about ophthalmic, retinal and even systemic diseases such as hypertension, diabetes, macular degeneration

More information

CS145: INTRODUCTION TO DATA MINING

CS145: INTRODUCTION TO DATA MINING CS145: INTRODUCTION TO DATA MINING 08: Classification Evaluation and Practical Issues Instructor: Yizhou Sun yzsun@cs.ucla.edu October 24, 2017 Learnt Prediction and Classification Methods Vector Data

More information

Natural Scene Sta,s,cs of Color and Range. Che- Chun Su, Lawrence K. Cormack, and Alan C. Bovik

Natural Scene Sta,s,cs of Color and Range. Che- Chun Su, Lawrence K. Cormack, and Alan C. Bovik Natural Scene Sta,s,cs of Color and Range Che- Chun Su, Lawrence K. Cormack, and Alan C. Bovik Mo,va,on Color and range/depth play important roles in natural scenes and human vision systems. Percep,on

More information

Feature Descriptors. CS 510 Lecture #21 April 29 th, 2013

Feature Descriptors. CS 510 Lecture #21 April 29 th, 2013 Feature Descriptors CS 510 Lecture #21 April 29 th, 2013 Programming Assignment #4 Due two weeks from today Any questions? How is it going? Where are we? We have two umbrella schemes for object recognition

More information

Ensemble- Based Characteriza4on of Uncertain Features Dennis McLaughlin, Rafal Wojcik

Ensemble- Based Characteriza4on of Uncertain Features Dennis McLaughlin, Rafal Wojcik Ensemble- Based Characteriza4on of Uncertain Features Dennis McLaughlin, Rafal Wojcik Hydrology TRMM TMI/PR satellite rainfall Neuroscience - - MRI Medicine - - CAT Geophysics Seismic Material tes4ng Laser

More information

Detection of Bone Fracture using Image Processing Methods

Detection of Bone Fracture using Image Processing Methods Detection of Bone Fracture using Image Processing Methods E Susmitha, M.Tech Student, Susmithasrinivas3@gmail.com Mr. K. Bhaskar Assistant Professor bhasi.adc@gmail.com MVR college of engineering and Technology

More information

SIFT (Scale Invariant Feature Transform) descriptor

SIFT (Scale Invariant Feature Transform) descriptor Local descriptors SIFT (Scale Invariant Feature Transform) descriptor SIFT keypoints at loca;on xy and scale σ have been obtained according to a procedure that guarantees illumina;on and scale invance.

More information

Segmentation and Modeling of the Spinal Cord for Reality-based Surgical Simulator

Segmentation and Modeling of the Spinal Cord for Reality-based Surgical Simulator Segmentation and Modeling of the Spinal Cord for Reality-based Surgical Simulator Li X.C.,, Chui C. K.,, and Ong S. H.,* Dept. of Electrical and Computer Engineering Dept. of Mechanical Engineering, National

More information

Webpage: Volume 3, Issue V, May 2015 eissn:

Webpage:   Volume 3, Issue V, May 2015 eissn: Morphological Image Processing of MRI Brain Tumor Images Using MATLAB Sarla Yadav 1, Parul Yadav 2 and Dinesh K. Atal 3 Department of Biomedical Engineering Deenbandhu Chhotu Ram University of Science

More information

Spatial Information Based Image Classification Using Support Vector Machine

Spatial Information Based Image Classification Using Support Vector Machine Spatial Information Based Image Classification Using Support Vector Machine P.Jeevitha, Dr. P. Ganesh Kumar PG Scholar, Dept of IT, Regional Centre of Anna University, Coimbatore, India. Assistant Professor,

More information

Energy- Aware Time Change Detec4on Using Synthe4c Aperture Radar On High- Performance Heterogeneous Architectures: A DDDAS Approach

Energy- Aware Time Change Detec4on Using Synthe4c Aperture Radar On High- Performance Heterogeneous Architectures: A DDDAS Approach Energy- Aware Time Change Detec4on Using Synthe4c Aperture Radar On High- Performance Heterogeneous Architectures: A DDDAS Approach Sanjay Ranka (PI) Sartaj Sahni (Co- PI) Mark Schmalz (Co- PI) University

More information

Classification of Protein Crystallization Imagery

Classification of Protein Crystallization Imagery Classification of Protein Crystallization Imagery Xiaoqing Zhu, Shaohua Sun, Samuel Cheng Stanford University Marshall Bern Palo Alto Research Center September 2004, EMBC 04 Outline Background X-ray crystallography

More information

Decision Trees, Random Forests and Random Ferns. Peter Kovesi

Decision Trees, Random Forests and Random Ferns. Peter Kovesi Decision Trees, Random Forests and Random Ferns Peter Kovesi What do I want to do? Take an image. Iden9fy the dis9nct regions of stuff in the image. Mark the boundaries of these regions. Recognize and

More information

More Learning. Ensembles Bayes Rule Neural Nets K-means Clustering EM Clustering WEKA

More Learning. Ensembles Bayes Rule Neural Nets K-means Clustering EM Clustering WEKA More Learning Ensembles Bayes Rule Neural Nets K-means Clustering EM Clustering WEKA 1 Ensembles An ensemble is a set of classifiers whose combined results give the final decision. test feature vector

More information

Practice Exam Sample Solutions

Practice Exam Sample Solutions CS 675 Computer Vision Instructor: Marc Pomplun Practice Exam Sample Solutions Note that in the actual exam, no calculators, no books, and no notes allowed. Question 1: out of points Question 2: out of

More information

Hybrid Approach for MRI Human Head Scans Classification using HTT based SFTA Texture Feature Extraction Technique

Hybrid Approach for MRI Human Head Scans Classification using HTT based SFTA Texture Feature Extraction Technique Volume 118 No. 17 2018, 691-701 ISSN: 1311-8080 (printed version); ISSN: 1314-3395 (on-line version) url: http://www.ijpam.eu ijpam.eu Hybrid Approach for MRI Human Head Scans Classification using HTT

More information

Generalized Hough Transform, line fitting

Generalized Hough Transform, line fitting Generalized Hough Transform, line fitting Introduction to Computer Vision CSE 152 Lecture 11-a Announcements Assignment 2: Due today Midterm: Thursday, May 10 in class Non-maximum suppression For every

More information

Region-based Segmentation

Region-based Segmentation Region-based Segmentation Image Segmentation Group similar components (such as, pixels in an image, image frames in a video) to obtain a compact representation. Applications: Finding tumors, veins, etc.

More information

INF4820 Algorithms for AI and NLP. Evaluating Classifiers Clustering

INF4820 Algorithms for AI and NLP. Evaluating Classifiers Clustering INF4820 Algorithms for AI and NLP Evaluating Classifiers Clustering Erik Velldal & Stephan Oepen Language Technology Group (LTG) September 23, 2015 Agenda Last week Supervised vs unsupervised learning.

More information

Lecture 13: Tracking mo3on features op3cal flow

Lecture 13: Tracking mo3on features op3cal flow Lecture 13: Tracking mo3on features op3cal flow Professor Fei- Fei Li Stanford Vision Lab Lecture 13-1! What we will learn today? Introduc3on Op3cal flow Feature tracking Applica3ons (Problem Set 3 (Q1))

More information

Ar#ficial Intelligence

Ar#ficial Intelligence Ar#ficial Intelligence Advanced Searching Prof Alexiei Dingli Gene#c Algorithms Charles Darwin Genetic Algorithms are good at taking large, potentially huge search spaces and navigating them, looking for

More information

TOWARDS AUTOMATED DEM GENERATION FROM HIGH RESOLUTION STEREO SATELLITE IMAGES

TOWARDS AUTOMATED DEM GENERATION FROM HIGH RESOLUTION STEREO SATELLITE IMAGES TOWARDS AUTOMATED DEM GENERATION FROM HIGH RESOLUTION STEREO SATELLITE IMAGES Pablo d'angelo, Manfred Lehner, Thomas Krauss, Danielle Hoja and Peter Reinartz German Aerospace Center (DLR), Remote Sensing

More information

Graph Based Classification of MRI Data Based on the Ventricular System

Graph Based Classification of MRI Data Based on the Ventricular System 1 Graph Based Classification of MRI Data Based on the Ventricular System S. Seth Long and Lawrence B. Holder Washington State University Abstract The ventricular system inside the brain is known to enlarge

More information

Informa/on Retrieval. Text Search. CISC437/637, Lecture #23 Ben CartereAe. Consider a database consis/ng of long textual informa/on fields

Informa/on Retrieval. Text Search. CISC437/637, Lecture #23 Ben CartereAe. Consider a database consis/ng of long textual informa/on fields Informa/on Retrieval CISC437/637, Lecture #23 Ben CartereAe Copyright Ben CartereAe 1 Text Search Consider a database consis/ng of long textual informa/on fields News ar/cles, patents, web pages, books,

More information

Interna'onal Community for Open and Interoperable AR content and experiences

Interna'onal Community for Open and Interoperable AR content and experiences where professionals get the latest information about standards for AR Interna'onal Community for Open and Interoperable AR content and experiences Summary Report of Fourth Mee'ng Oct 24-25 2011 At- a-

More information

Monocular Head Pose Es0ma0on

Monocular Head Pose Es0ma0on ICIAR 2008 - Interna0onal Conference on Image Analysis and Recogni0on Monocular Head Pose Es0ma0on Pedro Mar0ns, Jorge Ba0sta Ins0tute for Systems and Robo0cs h>p://www.isr.uc.pt Department of Electrical

More information

Volume Illumination and Segmentation

Volume Illumination and Segmentation Volume Illumination and Segmentation Computer Animation and Visualisation Lecture 13 Institute for Perception, Action & Behaviour School of Informatics Overview Volume illumination Segmentation Volume

More information

Chapter 5: Outlier Detection

Chapter 5: Outlier Detection Ludwig-Maximilians-Universität München Institut für Informatik Lehr- und Forschungseinheit für Datenbanksysteme Knowledge Discovery in Databases SS 2016 Chapter 5: Outlier Detection Lecture: Prof. Dr.

More information

A (Very) Brief Introduc3on to Machine Learning and its Applica3on to Mobile Networks. David Meyer SP CTO and Chief Scien3st Brocade

A (Very) Brief Introduc3on to Machine Learning and its Applica3on to Mobile Networks. David Meyer SP CTO and Chief Scien3st Brocade A (Very) Brief Introduc3on to Machine Learning and its Applica3on to Mobile Networks David Meyer SP CTO and Chief Scien3st Brocade dmm@brocade.com Agenda Goals for this Talk Automa3on Con3nuum SoKware

More information

Learning from Data. COMP61011 : Machine Learning and Data Mining. Dr Gavin Brown Machine Learning and Op<miza<on Research Group

Learning from Data. COMP61011 : Machine Learning and Data Mining. Dr Gavin Brown Machine Learning and Op<miza<on Research Group Learning from Data COMP61011 : Machine Learning and Data Mining Dr Gavin Brown Machine Learning and Op

More information

Lecture 9: Hough Transform and Thresholding base Segmentation

Lecture 9: Hough Transform and Thresholding base Segmentation #1 Lecture 9: Hough Transform and Thresholding base Segmentation Saad Bedros sbedros@umn.edu Hough Transform Robust method to find a shape in an image Shape can be described in parametric form A voting

More information

Preliminary ACTL-SLOW Design in the ACS and OPC-UA context. G. Tos? (19/04/2016)

Preliminary ACTL-SLOW Design in the ACS and OPC-UA context. G. Tos? (19/04/2016) Preliminary ACTL-SLOW Design in the ACS and OPC-UA context G. Tos? (19/04/2016) Summary General Introduc?on to ACS Preliminary ACTL-SLOW proposed design Hardware device integra?on in ACS and ACTL- SLOW

More information

EE663 Image Processing Histogram Equalization I

EE663 Image Processing Histogram Equalization I EE663 Image Processing Histogram Equalization I Dr. Samir H. Abdul-Jauwad Electrical Engineering Department College of Engineering Sciences King Fahd University of Petroleum & Minerals Dhahran Saudi Arabia

More information

Lecture 14: Tracking mo3on features op3cal flow

Lecture 14: Tracking mo3on features op3cal flow Lecture 14: Tracking mo3on features op3cal flow Dr. Juan Carlos Niebles Stanford AI Lab Professor Fei- Fei Li Stanford Vision Lab Lecture 14-1 What we will learn today? Introduc3on Op3cal flow Feature

More information

Detec%ng Wildlife in Uncontrolled Outdoor Video using Convolu%onal Neural Networks

Detec%ng Wildlife in Uncontrolled Outdoor Video using Convolu%onal Neural Networks Detec%ng Wildlife in Uncontrolled Outdoor Video using Convolu%onal Neural Networks Connor Bowley *, Alicia Andes +, Susan Ellis-Felege +, Travis Desell * Department of Computer Science * Department of

More information

Fix- point engine in Z3. Krystof Hoder Nikolaj Bjorner Leonardo de Moura

Fix- point engine in Z3. Krystof Hoder Nikolaj Bjorner Leonardo de Moura μz Fix- point engine in Z3 Krystof Hoder Nikolaj Bjorner Leonardo de Moura Mo?va?on Horn EPR applica?ons (Datalog) Points- to analysis Security analysis Deduc?ve data- bases and knowledge bases (Yago)

More information

Spa$al Analysis and Modeling (GIST 4302/5302) Guofeng Cao Department of Geosciences Texas Tech University

Spa$al Analysis and Modeling (GIST 4302/5302) Guofeng Cao Department of Geosciences Texas Tech University Spa$al Analysis and Modeling (GIST 432/532) Guofeng Cao Department of Geosciences Texas Tech University Representa$on of Spa$al Data Representa$on of Spa$al Data Models Object- based model: treats the

More information

PhD in Computer And Control Engineering XXVII cycle. Torino February 27th, 2015.

PhD in Computer And Control Engineering XXVII cycle. Torino February 27th, 2015. PhD in Computer And Control Engineering XXVII cycle Torino February 27th, 2015. Parallel and reconfigurable systems are more and more used in a wide number of applica7ons and environments, ranging from

More information

COMP 9517 Computer Vision

COMP 9517 Computer Vision COMP 9517 Computer Vision Frequency Domain Techniques 1 Frequency Versus Spa

More information

L6: System design: behavior models

L6: System design: behavior models L6: System design: behavior models Limita6ons of func6onal decomposi6on Behavior models State diagrams Flow charts Data flow diagrams En6ty rela6onship diagrams Unified Modeling Language Capstone design

More information

TerraSwarm. A Machine Learning and Op0miza0on Toolkit for the Swarm. Ilge Akkaya, Shuhei Emoto, Edward A. Lee. University of California, Berkeley

TerraSwarm. A Machine Learning and Op0miza0on Toolkit for the Swarm. Ilge Akkaya, Shuhei Emoto, Edward A. Lee. University of California, Berkeley TerraSwarm A Machine Learning and Op0miza0on Toolkit for the Swarm Ilge Akkaya, Shuhei Emoto, Edward A. Lee University of California, Berkeley TerraSwarm Tools Telecon 17 November 2014 Sponsored by the

More information

CHAPTER 3 RETINAL OPTIC DISC SEGMENTATION

CHAPTER 3 RETINAL OPTIC DISC SEGMENTATION 60 CHAPTER 3 RETINAL OPTIC DISC SEGMENTATION 3.1 IMPORTANCE OF OPTIC DISC Ocular fundus images provide information about ophthalmic, retinal and even systemic diseases such as hypertension, diabetes, macular

More information

Lab 9. Julia Janicki. Introduction

Lab 9. Julia Janicki. Introduction Lab 9 Julia Janicki Introduction My goal for this project is to map a general land cover in the area of Alexandria in Egypt using supervised classification, specifically the Maximum Likelihood and Support

More information

CS443: Digital Imaging and Multimedia Perceptual Grouping Detecting Lines and Simple Curves

CS443: Digital Imaging and Multimedia Perceptual Grouping Detecting Lines and Simple Curves CS443: Digital Imaging and Multimedia Perceptual Grouping Detecting Lines and Simple Curves Spring 2008 Ahmed Elgammal Dept. of Computer Science Rutgers University Outlines Perceptual Grouping and Segmentation

More information