Towards Improved Epilepsia Diagnosis by Unsupervised Segmentation of Neuropathology Tissue Sections using Ripley s-ˆl Features
|
|
- Anna Kerrie Cunningham
- 5 years ago
- Views:
Transcription
1 Towards Improved Epilepsia Diagnosis by Unsupervised Segmentation of Neuropathology Tissue Sections using Ripley s-ˆl Features Timm Schoening 1, Volkmar H. Hans 2, Tim W. Nattkemper 1 1 Biodata Mining Group, Faculty of Technology, Bielefeld University 2 Institute of Neuropathology, Evangelisches Krankenhaus Bielefeld, Germany tschoeni@cebitec.uni-bielefeld.de Abstract. The analysis of architectural features in neural tissue sections and the identification of distinct regions is challenging for computer aided diagnosis (CAD) in neuropathology. Due to the difficulty of locating a tissue s origin and alignment as well as the vast variety of structures within such images an orientation independent (i. e. rotation invariant) approach for tissue region segmentation has to be found to encode the structural features of neural layer architecture in the tissue. We propose to apply the Ripley s-ˆl function, originating from the field of plant ecology, to compute feature vectors encoding the spatial statistics of point patterns described by selectively stained cells. Combining the Ripley s ˆL features with unsupervised clustering enables a segmentation of tissue sections into neuropathological areas. 1 Introduction Visual inspection of microscopy images from neural tissue sections taken from human brains immediately show a variety of structure orientation and distribution of neural cells (Fig. 1a). Depending on the tissue s origin, the position and direction of the cut and the selected part of the tissue to be captured, several input factors determine the initial data to be analyzed. The images analyzed here contained captions of the cortical layers of the human brain. These layers are indexed by roman numbers. Beginning at the outside with layer I which is an almost empty region, five layers of different internal structure follow, ending with layer VI where the amount of neurons faints towards the white matter (Fig. 1b). By just looking at the images, be it the original microscopy outcome or a binarized or centroided one, it is possible to distinguish some major structures although it is a subjective task. Hence the challenge is to enable a computer to identify similar regions, even by means of minor dissimilarities, making it possible to get an objective, reproducible and understandable identification of tissue structures and distortions in the neural architecture. To gain an unbiased insight to the data, we chose to apply the Ripley s-ˆl function, which has only recently been introduced to the field of microscopy [1].
2 2 Materials and Methods 2.1 Imaging Protocol Towards Improved Epilepsia Diagnosis 45 Diagnostic samples were derived from neurosurgical resections at the Bethel Epilepsy Center, Bielefeld, Germany for therapeutic purposes, routinely processed for formalin fixation and paraffin embedded. Four µm thick microtome sections were immunostained using an antibody against NeuN, a marker labeling almost all mature cortical neurons. Positive staining of cells rendered them brown, leaving surrounding structures with a faint blue counterstain. Archival slides of 19 cases from the years 2009 and 2010 were selected for representing largely normal lateral temporal isocortex cut perpendicular to the brain surface. Slides were analyzed on a Zeiss Axioskop 2 plus microscope with a Zeiss Achroplan 2.5x/0.07 lens. After manually focusing and automated background correction, pixels, 24 bit, true color RGB pictures were taken at standardized 3200 K light temperature in TIF format using Zeiss AxioVision 3.1 software and a Zeiss AxioCam HRc digital color camera (Carl Zeiss AG, Oberkochen, Germany). All images are fully anonymized and our work did not influence the diagnostic process or the patient s treatment. 2.2 Segmentation The Ripley s- ˆK function [2] is defined for point patterns, so nuclei segmentation was applied to all images using a color channel based algorithm, that takes the average of the three channels of the RGB image at each pixel and compares this value to a given thresholding parameter. All values below this threshold were treated as background, all values above or equal were thought of as neural cells. The binary image for different thresholds was visually compared to the estimated result and a threshold value of about 70 has shown good outcomes. To get a point pattern, the binarized image was segmented, i.e. all connected pixels were combined to a cell region and the centroid of this region was computed as the cells representative point (RP). All RPs together create the nuclei point pattern of an image or an image region. 2.3 Ripleys- ˆL Function The original Ripley function was developed as the Ripley s- ˆK function which analyses the distribution of n points in a given area A. It centers an imaginary circle of a given radius at each point of the observed pattern and counts the number of other points found within this circle. This number is set into relation to the expected amount of points within the circle determined by the size of the complete region ( study region A) and the total number of points (n) and then summed up for all points i. The ˆK value ˆK(r) = A n 2 n n i=1 j=1 δ ij (r) C i (r) (1)
3 46 Schoening et al. is computed for all radii within a given interval (i.e. from r = 1 to r = r max ), with δ in the numerator is 1 for point distances d(i,j) < r, else it is 0. The denominator is the fraction of the circle area with radius r at point i within the study region so the point counts for points i close to the study region borders are adjusted [3]. A drawback of the ˆK function are its fast growing values for increasing radii. Therefore the ˆL function was introduced which is a normalized representation and has values around zero ˆL(r) = ˆK(r) π r (2) ˆL values above zero for a given radius indicate more, values below zero indicate less points within this radius than expected. 2.4 Confidence Envelopes To evaluate the significance of an observed point pattern, a reference point pattern is required. Therefore usually the assumed pattern creation process is simulated to create a set of further point patterns [4]. There is a variety of possible point processes, the easiest of which is complete spatial randomness which was used here. We simulated a similar study region and randomly distributed the same amount of points as in the original pattern within this region. We then again calculated the ˆL function and repeated this procedure nineteen times to yield 95 percent confidence envelopes by taking the highest and lowest calculated values for each radius as envelope limits. 2.5 Feature Set Each captured pixel image was cut into overlapping tiles. Each tile was pixel and the overlap in each direction was 140 pixels creating a set of tiles. Each tile was represented by one Ripley s-ˆl feature vector, yielding 10,324 feature vectors per image. To avoid edge effects the tiles had to be completely inside the image, leaving a 70 pixel border at each side which was not analyzed. We then computed standard point pattern features like density (λ), mean point distance ( d) and deviation of mean distance (σ). The density gave good approximations for the underlying brain structure as can be expected. But being a one-dimensional measure, it was not able to identify semantic differences between the layers, e.g. layers I and VI (Fig. 1, B) both feature low density but resemble the far inner/outer layer and show distinctive structures at the transitions towards their neighboring layers. To compute Ripley s-ˆl based features, we computed the ˆL function (with r max = 20) on each tile s point pattern, so A = in (1), and calculated features from the resulting ˆL function like the amount of null points (N), area above (I a ), below (I b ) and outside (I o ) confidence envelopes. In order to achieve a compact representation of the ˆL function shape, we approximated the ˆL-function by a polynomial of
4 Towards Improved Epilepsia Diagnosis 47 degree m and used the polynomial coefficients as features. However in our iniˆ functions numerical tial experiments we found that we had to include all the L values to get a more comprehensive description of its slope. Since we needed to take the confidence envelopes into account here, we calculated a shifted version ˆ function, that is zero for all radii at which the function is within the of the L confidence envelope and the difference between the function and the envelope otherwise. As a result of several experiments we found, that a joint feature vector v, containing density (λ) and all numerical values of the original as well as ˆ function, produced the best results and so we joined those features the shifted L to a 41-dim feature space. 2.6 Clustering, Segmentation and Visualization The feature vectors from all tiles of all images were fused to one training set (19 10,324 = 196,156 items). Each vector was normalized to v = 1 before clustering. Batch k-means was applied to fit k=20 prototype vectors vα,α=1..20 to regions of high feature vector density. The clustering was stopped if at most 50 iterations were performed or less than 0.1 % of the assignments of feature vectors to their best matching unit (BMU) changed during one iteration. Afterwards each pixel was mapped to its BMU to achieve a first segmentation i.e. Fig. 1. A: an original input image, beginning on the left with white matter, then layers VI to I, attached to the right is the next cerebral gyrus, again beginning with layer I and II; B: a hand labeled image with highlighted regions and an arbitrary color coding; C: The original image as a background with a density overlay; D: The clustering result where less occupied regions are dark blue whereas the filled layers are purple and pink at sharp edges towards empty regions.
5 48 Schoening et al. pixel labeling result. To visualize and evaluate this segmentation result, each prototype v α was assigned to a RGB color (r,g,b) α. To preserve topology in the mapping from the feature to the color space, we assigned RGB components using projections of the prototype vectors onto the eigenvectors belonging to the three highest eigenvalues of the feature set (Fig. 1d). We reran the experiments with the same settings to confirm the cluster results and found that small differences can occur due to the randomization effects during confidence envelope calculation. 3 Results The segmented result (Fig. 1d) shows a layered structure comparable to the density image (Fig. 1c). However the color scale compared to density of the prototypes shows that the prototypes describe a manifold of a higher dimension i.e. the features encode spatial features more complex than density. At least three eigenvectors, belonging to the three highest eigenvalues are necessary to encode 95 percent of the data variance. The less occupied regions are separated into the white matter and layer I. Layers II to VI show similar results as both feature pattern distributions of moderate to high density and are only discerned by their relative position to each other. The resulting structure shows good overlap with the hand labeled image (Fig. 1, B) although there are differences especially for layers V and VI where the subjective coloring shows a smaller region V but the data driven approach assigns more tiles to the comparable region. 4 Discussion The outcome of our proposed method used as an initial attempt is encouraging for further optimization. Beginning with binarization and centroid finding, moving on to feature selection and normalization, ending up with clustering and result visualization, every component of this project could be improved. On the other hand every part of it is simple and easily understandable which is important to avoid explanation problems of black-box solutions. Our results are promising to provide in the future a computer aided tool for visualizing subtle developmental brain abnormalities associated with human epilepsy. References 1. Mattfeld T, Eckel S, Fleischer F, et al. Statistical analysis of labelling patterns of mammary carcinoma cell nuclei on histological sections. J Microscopy. 2009;235: Ripley BD. The second-order analysis of stationary point processes. J Appl Prob. 1976;13: Goreaud F, Pelissier R. On explicit formulas of edge effect correction for Ripley s K-function. J Veget Sci. 1999;10: Wiegand T, Moloney KA. Rings, circles and null-models for point pattern analysis in ecology. Oikos. 2004;104:
SUPPLEMENTARY FILE S1: 3D AIRWAY TUBE RECONSTRUCTION AND CELL-BASED MECHANICAL MODEL. RELATED TO FIGURE 1, FIGURE 7, AND STAR METHODS.
SUPPLEMENTARY FILE S1: 3D AIRWAY TUBE RECONSTRUCTION AND CELL-BASED MECHANICAL MODEL. RELATED TO FIGURE 1, FIGURE 7, AND STAR METHODS. 1. 3D AIRWAY TUBE RECONSTRUCTION. RELATED TO FIGURE 1 AND STAR METHODS
More informationDetection of Focal Cortical Dysplasia Lesions in MRI using Textural Features
Detection of Focal Cortical Dysplasia Lesions in MRI using Textural Features Christian Loyek 1, Friedrich G. Woermann 2, Tim W. Nattkemper 1 1 Applied Neuroinformatics, Faculty of Technology, Bielefeld
More informationChapter 4. Clustering Core Atoms by Location
Chapter 4. Clustering Core Atoms by Location In this chapter, a process for sampling core atoms in space is developed, so that the analytic techniques in section 3C can be applied to local collections
More informationEE 584 MACHINE VISION
EE 584 MACHINE VISION Binary Images Analysis Geometrical & Topological Properties Connectedness Binary Algorithms Morphology Binary Images Binary (two-valued; black/white) images gives better efficiency
More informationSpectral Classification
Spectral Classification Spectral Classification Supervised versus Unsupervised Classification n Unsupervised Classes are determined by the computer. Also referred to as clustering n Supervised Classes
More informationClustering CS 550: Machine Learning
Clustering CS 550: Machine Learning This slide set mainly uses the slides given in the following links: http://www-users.cs.umn.edu/~kumar/dmbook/ch8.pdf http://www-users.cs.umn.edu/~kumar/dmbook/dmslides/chap8_basic_cluster_analysis.pdf
More informationOnline Pattern Recognition in Multivariate Data Streams using Unsupervised Learning
Online Pattern Recognition in Multivariate Data Streams using Unsupervised Learning Devina Desai ddevina1@csee.umbc.edu Tim Oates oates@csee.umbc.edu Vishal Shanbhag vshan1@csee.umbc.edu Machine Learning
More informationIMAGE SEGMENTATION. Václav Hlaváč
IMAGE SEGMENTATION Václav Hlaváč Czech Technical University in Prague Faculty of Electrical Engineering, Department of Cybernetics Center for Machine Perception http://cmp.felk.cvut.cz/ hlavac, hlavac@fel.cvut.cz
More informationFast and accurate automated cell boundary determination for fluorescence microscopy
Fast and accurate automated cell boundary determination for fluorescence microscopy Stephen Hugo Arce, Pei-Hsun Wu &, and Yiider Tseng Department of Chemical Engineering, University of Florida and National
More informationCLASSIFICATION WITH RADIAL BASIS AND PROBABILISTIC NEURAL NETWORKS
CLASSIFICATION WITH RADIAL BASIS AND PROBABILISTIC NEURAL NETWORKS CHAPTER 4 CLASSIFICATION WITH RADIAL BASIS AND PROBABILISTIC NEURAL NETWORKS 4.1 Introduction Optical character recognition is one of
More informationImage Segmentation. Shengnan Wang
Image Segmentation Shengnan Wang shengnan@cs.wisc.edu Contents I. Introduction to Segmentation II. Mean Shift Theory 1. What is Mean Shift? 2. Density Estimation Methods 3. Deriving the Mean Shift 4. Mean
More informationThe organization of the human cerebral cortex estimated by intrinsic functional connectivity
1 The organization of the human cerebral cortex estimated by intrinsic functional connectivity Journal: Journal of Neurophysiology Author: B. T. Thomas Yeo, et al Link: https://www.ncbi.nlm.nih.gov/pubmed/21653723
More informationMachine Learning in Biology
Università degli studi di Padova Machine Learning in Biology Luca Silvestrin (Dottorando, XXIII ciclo) Supervised learning Contents Class-conditional probability density Linear and quadratic discriminant
More informationCS443: Digital Imaging and Multimedia Binary Image Analysis. Spring 2008 Ahmed Elgammal Dept. of Computer Science Rutgers University
CS443: Digital Imaging and Multimedia Binary Image Analysis Spring 2008 Ahmed Elgammal Dept. of Computer Science Rutgers University Outlines A Simple Machine Vision System Image segmentation by thresholding
More informationBiometrics Technology: Image Processing & Pattern Recognition (by Dr. Dickson Tong)
Biometrics Technology: Image Processing & Pattern Recognition (by Dr. Dickson Tong) References: [1] http://homepages.inf.ed.ac.uk/rbf/hipr2/index.htm [2] http://www.cs.wisc.edu/~dyer/cs540/notes/vision.html
More informationMotion Estimation. There are three main types (or applications) of motion estimation:
Members: D91922016 朱威達 R93922010 林聖凱 R93922044 謝俊瑋 Motion Estimation There are three main types (or applications) of motion estimation: Parametric motion (image alignment) The main idea of parametric motion
More informationDigital Image Processing
Digital Image Processing Part 9: Representation and Description AASS Learning Systems Lab, Dep. Teknik Room T1209 (Fr, 11-12 o'clock) achim.lilienthal@oru.se Course Book Chapter 11 2011-05-17 Contents
More informationClustering & Dimensionality Reduction. 273A Intro Machine Learning
Clustering & Dimensionality Reduction 273A Intro Machine Learning What is Unsupervised Learning? In supervised learning we were given attributes & targets (e.g. class labels). In unsupervised learning
More informationCSE152a Computer Vision Assignment 2 WI14 Instructor: Prof. David Kriegman. Revision 1
CSE152a Computer Vision Assignment 2 WI14 Instructor: Prof. David Kriegman. Revision 1 Instructions: This assignment should be solved, and written up in groups of 2. Work alone only if you can not find
More informationArtifacts and Textured Region Detection
Artifacts and Textured Region Detection 1 Vishal Bangard ECE 738 - Spring 2003 I. INTRODUCTION A lot of transformations, when applied to images, lead to the development of various artifacts in them. In
More informationEmotion Classification
Emotion Classification Shai Savir 038052395 Gil Sadeh 026511469 1. Abstract Automated facial expression recognition has received increased attention over the past two decades. Facial expressions convey
More informationSimplified model of ray propagation and outcoupling in TFs.
Supplementary Figure 1 Simplified model of ray propagation and outcoupling in TFs. After total reflection at the core boundary with an angle α, a ray entering the taper (blue line) hits the taper sidewalls
More informationConnected components - 1
Connected Components Basic definitions Connectivity, Adjacency, Connected Components Background/Foreground, Boundaries Run-length encoding Component Labeling Recursive algorithm Two-scan algorithm Chain
More informationA Generalized Method to Solve Text-Based CAPTCHAs
A Generalized Method to Solve Text-Based CAPTCHAs Jason Ma, Bilal Badaoui, Emile Chamoun December 11, 2009 1 Abstract We present work in progress on the automated solving of text-based CAPTCHAs. Our method
More informationIntroduction to Trajectory Clustering. By YONGLI ZHANG
Introduction to Trajectory Clustering By YONGLI ZHANG Outline 1. Problem Definition 2. Clustering Methods for Trajectory data 3. Model-based Trajectory Clustering 4. Applications 5. Conclusions 1 Problem
More informationCluster Analysis. Mu-Chun Su. Department of Computer Science and Information Engineering National Central University 2003/3/11 1
Cluster Analysis Mu-Chun Su Department of Computer Science and Information Engineering National Central University 2003/3/11 1 Introduction Cluster analysis is the formal study of algorithms and methods
More informationDetection of Ductus in Mammary Gland Tissue by Computer Vision
Detection of Ductus in Mammary Gland Tissue by Computer Vision Katarina Mele and Aleš Leonardis Computer Vision Laboratory Faculty of Computer and Information Science University of Ljubljana Tržaška 25,
More informationNetwork Snakes for the Segmentation of Adjacent Cells in Confocal Images
Network Snakes for the Segmentation of Adjacent Cells in Confocal Images Matthias Butenuth 1 and Fritz Jetzek 2 1 Institut für Photogrammetrie und GeoInformation, Leibniz Universität Hannover, 30167 Hannover
More informationImage analysis in IHC - overview, considerations and applications
Image analysis in IHC - overview, considerations and applications Workshop in Diagnostic Immunohistochemistry Oud St. Jan/ Old St. John Brugge (Bruges), Belgium June 13th 15nd 2018 Rasmus Røge, MD, NordiQC
More informationA Neural Network Architecture for Automatic Segmentation of Fluorescence Micrographs
A Network Architecture for Automatic Segmentation of Fluorescence Micrographs Tim W. Nattkemper 1, Heiko Wersing 1 Walter Schubert 2;3, Helge Ritter 1 1 Neuroinformatics Group, University of Bielefeld,
More informationRange Imaging Through Triangulation. Range Imaging Through Triangulation. Range Imaging Through Triangulation. Range Imaging Through Triangulation
Obviously, this is a very slow process and not suitable for dynamic scenes. To speed things up, we can use a laser that projects a vertical line of light onto the scene. This laser rotates around its vertical
More informationTutorial for cell nucleus intensity clustering on histochemically stained tissue microarrays.
Tutorial for cell nucleus intensity clustering on histochemically stained tissue microarrays. http://www.nexus.ethz.ch/ -> Software -> TMARKER The plugin Intensity Clustering starts on step 7. Steps 1-6
More informationCoE4TN4 Image Processing
CoE4TN4 Image Processing Chapter 11 Image Representation & Description Image Representation & Description After an image is segmented into regions, the regions are represented and described in a form suitable
More informationAlgorithm User Guide:
Algorithm User Guide: Membrane Quantification Use the Aperio algorithms to adjust (tune) the parameters until the quantitative results are sufficiently accurate for the purpose for which you intend to
More information09/11/2017. Morphological image processing. Morphological image processing. Morphological image processing. Morphological image processing (binary)
Towards image analysis Goal: Describe the contents of an image, distinguishing meaningful information from irrelevant one. Perform suitable transformations of images so as to make explicit particular shape
More information8/3/2017. Contour Assessment for Quality Assurance and Data Mining. Objective. Outline. Tom Purdie, PhD, MCCPM
Contour Assessment for Quality Assurance and Data Mining Tom Purdie, PhD, MCCPM Objective Understand the state-of-the-art in contour assessment for quality assurance including data mining-based techniques
More informationAutomated Particle Size & Shape Analysis System
Biovis PSA2000 Automated Particle Size & Shape Analysis System Biovis PSA2000 is an automated imaging system used to detect, characterize, categorize and report, the individual and cumulative particle
More informationContents. Supplementary Information. Detection and Segmentation of Cell Nuclei in Virtual Microscopy Images: A Minimum-Model Approach
Supplementary Information Detection and Segmentation of Cell Nuclei in Virtual Microscopy Images: A Minimum-Model Approach Stephan Wienert 1,2, Daniel Heim 2, Kai Saeger 2, Albrecht Stenzinger 3, Michael
More informationLecture 10: Semantic Segmentation and Clustering
Lecture 10: Semantic Segmentation and Clustering Vineet Kosaraju, Davy Ragland, Adrien Truong, Effie Nehoran, Maneekwan Toyungyernsub Department of Computer Science Stanford University Stanford, CA 94305
More informationOut-of-sample extension of diffusion maps in a computer-aided diagnosis system. Application to breast cancer virtual slide images.
Out-of-sample extension of diffusion maps in a computer-aided diagnosis system. Application to breast cancer virtual slide images. Philippe BELHOMME Myriam OGER Jean-Jacques MICHELS Benoit PLANCOULAINE
More informationSupporting Information. High-Throughput, Algorithmic Determination of Nanoparticle Structure From Electron Microscopy Images
Supporting Information High-Throughput, Algorithmic Determination of Nanoparticle Structure From Electron Microscopy Images Christine R. Laramy, 1, Keith A. Brown, 2, Matthew N. O Brien, 2 and Chad. A.
More informationUnsupervised Learning
Unsupervised Learning Learning without a teacher No targets for the outputs Networks which discover patterns, correlations, etc. in the input data This is a self organisation Self organising networks An
More informationPractice 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 informationToward Automated Cancer Diagnosis: An Interactive System for Cell Feature Extraction
Toward Automated Cancer Diagnosis: An Interactive System for Cell Feature Extraction Nick Street Computer Sciences Department University of Wisconsin-Madison street@cs.wisc.edu Abstract Oncologists at
More informationAutomatically Improving 3D Neuron Segmentations for Expansion Microscopy Connectomics. by Albert Gerovitch
Automatically Improving 3D Neuron Segmentations for Expansion Microscopy Connectomics by Albert Gerovitch 1 Abstract Understanding the geometry of neurons and their connections is key to comprehending
More informationUNIVERSITY OF CALIFORNIA RIVERSIDE MAGIC CAMERA. A project report submitted in partial satisfaction of the requirements of the degree of
UNIVERSITY OF CALIFORNIA RIVERSIDE MAGIC CAMERA A project report submitted in partial satisfaction of the requirements of the degree of Master of Science in Computer Science by Adam Meadows June 2006 Project
More information3D RECONSTRUCTION OF BRAIN TISSUE
3D RECONSTRUCTION OF BRAIN TISSUE Hylke Buisman, Manuel Gomez-Rodriguez, Saeed Hassanpour {hbuisman, manuelgr, saeedhp}@stanford.edu Department of Computer Science, Department of Electrical Engineering
More informationArtificial Neural Networks Unsupervised learning: SOM
Artificial Neural Networks Unsupervised learning: SOM 01001110 01100101 01110101 01110010 01101111 01101110 01101111 01110110 01100001 00100000 01110011 01101011 01110101 01110000 01101001 01101110 01100001
More informationA Hillclimbing Approach to Image Mosaics
A Hillclimbing Approach to Image Mosaics Chris Allen Faculty Sponsor: Kenny Hunt, Department of Computer Science ABSTRACT This paper presents a hillclimbing approach to image mosaic creation. Our approach
More informationFeature Detectors and Descriptors: Corners, Lines, etc.
Feature Detectors and Descriptors: Corners, Lines, etc. Edges vs. Corners Edges = maxima in intensity gradient Edges vs. Corners Corners = lots of variation in direction of gradient in a small neighborhood
More informationThe Anatomical Equivalence Class Formulation and its Application to Shape-based Computational Neuroanatomy
The Anatomical Equivalence Class Formulation and its Application to Shape-based Computational Neuroanatomy Sokratis K. Makrogiannis, PhD From post-doctoral research at SBIA lab, Department of Radiology,
More informationClustering algorithms and autoencoders for anomaly detection
Clustering algorithms and autoencoders for anomaly detection Alessia Saggio Lunch Seminars and Journal Clubs Université catholique de Louvain, Belgium 3rd March 2017 a Outline Introduction Clustering algorithms
More informationLook Ahead to Get Ahead
Microscopy from Carl Zeiss Archive 3D Deconvolution Z-stack 3D Deconvolution Look Ahead to Get Ahead Mark & Find Multichannel Timelapse Greater brilliance and resolution More convenient and precise than
More informationlayers in a raster model
layers in a raster model Layer 1 Layer 2 layers in an vector-based model (1) Layer 2 Layer 1 layers in an vector-based model (2) raster versus vector data model Raster model Vector model Simple data structure
More information[7.3, EA], [9.1, CMB]
K-means Clustering Ke Chen Reading: [7.3, EA], [9.1, CMB] Outline Introduction K-means Algorithm Example How K-means partitions? K-means Demo Relevant Issues Application: Cell Neulei Detection Summary
More informationUsing Machine Learning for Classification of Cancer Cells
Using Machine Learning for Classification of Cancer Cells Camille Biscarrat University of California, Berkeley I Introduction Cell screening is a commonly used technique in the development of new drugs.
More informationRegion-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 informationClustering in Data Mining
Clustering in Data Mining Classification Vs Clustering When the distribution is based on a single parameter and that parameter is known for each object, it is called classification. E.g. Children, young,
More informationChapter 7: Competitive learning, clustering, and self-organizing maps
Chapter 7: Competitive learning, clustering, and self-organizing maps António R. C. Paiva EEL 6814 Spring 2008 Outline Competitive learning Clustering Self-Organizing Maps What is competition in neural
More informationK-Means Clustering Using Localized Histogram Analysis
K-Means Clustering Using Localized Histogram Analysis Michael Bryson University of South Carolina, Department of Computer Science Columbia, SC brysonm@cse.sc.edu Abstract. The first step required for many
More informationMULTI-REGION SEGMENTATION
MULTI-REGION SEGMENTATION USING GRAPH-CUTS Johannes Ulén Abstract This project deals with multi-region segmenation using graph-cuts and is mainly based on a paper by Delong and Boykov [1]. The difference
More informationData Mining 4. Cluster Analysis
Data Mining 4. Cluster Analysis 4.5 Spring 2010 Instructor: Dr. Masoud Yaghini Introduction DBSCAN Algorithm OPTICS Algorithm DENCLUE Algorithm References Outline Introduction Introduction Density-based
More informationSISCOM (Subtraction Ictal SPECT CO-registered to MRI)
SISCOM (Subtraction Ictal SPECT CO-registered to MRI) Introduction A method for advanced imaging of epilepsy patients has been developed with Analyze at the Mayo Foundation which uses a combination of
More informationMulti prototype fuzzy pattern matching for handwritten character recognition
Multi prototype fuzzy pattern matching for handwritten character recognition MILIND E. RANE, DHABE P. S AND J. B. PATIL Dept. of Electronics and Computer, R.C. Patel Institute of Technology, Shirpur, Dist.
More informationCanny Edge Based Self-localization of a RoboCup Middle-sized League Robot
Canny Edge Based Self-localization of a RoboCup Middle-sized League Robot Yoichi Nakaguro Sirindhorn International Institute of Technology, Thammasat University P.O. Box 22, Thammasat-Rangsit Post Office,
More informationCounting Particles or Cells Using IMAQ Vision
Application Note 107 Counting Particles or Cells Using IMAQ Vision John Hanks Introduction To count objects, you use a common image processing technique called particle analysis, often referred to as blob
More informationRobust PDF Table Locator
Robust PDF Table Locator December 17, 2016 1 Introduction Data scientists rely on an abundance of tabular data stored in easy-to-machine-read formats like.csv files. Unfortunately, most government records
More informationCrop Counting and Metrics Tutorial
Crop Counting and Metrics Tutorial The ENVI Crop Science platform contains remote sensing analytic tools for precision agriculture and agronomy. In this tutorial you will go through a typical workflow
More informationUnsupervised Learning : Clustering
Unsupervised Learning : Clustering Things to be Addressed Traditional Learning Models. Cluster Analysis K-means Clustering Algorithm Drawbacks of traditional clustering algorithms. Clustering as a complex
More informationChapter 11 Representation & Description
Chain Codes Chain codes are used to represent a boundary by a connected sequence of straight-line segments of specified length and direction. The direction of each segment is coded by using a numbering
More informationGlobal 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 informationShape-based Diffeomorphic Registration on Hippocampal Surfaces Using Beltrami Holomorphic Flow
Shape-based Diffeomorphic Registration on Hippocampal Surfaces Using Beltrami Holomorphic Flow Abstract. Finding meaningful 1-1 correspondences between hippocampal (HP) surfaces is an important but difficult
More information6. Dicretization methods 6.1 The purpose of discretization
6. Dicretization methods 6.1 The purpose of discretization Often data are given in the form of continuous values. If their number is huge, model building for such data can be difficult. Moreover, many
More informationA Keypoint Descriptor Inspired by Retinal Computation
A Keypoint Descriptor Inspired by Retinal Computation Bongsoo Suh, Sungjoon Choi, Han Lee Stanford University {bssuh,sungjoonchoi,hanlee}@stanford.edu Abstract. The main goal of our project is to implement
More informationCommunity Detection. Jian Pei: CMPT 741/459 Clustering (1) 2
Clustering Community Detection http://image.slidesharecdn.com/communitydetectionitilecturejune0-0609559-phpapp0/95/community-detection-in-social-media--78.jpg?cb=3087368 Jian Pei: CMPT 74/459 Clustering
More informationcoding of various parts showing different features, the possibility of rotation or of hiding covering parts of the object's surface to gain an insight
Three-Dimensional Object Reconstruction from Layered Spatial Data Michael Dangl and Robert Sablatnig Vienna University of Technology, Institute of Computer Aided Automation, Pattern Recognition and Image
More informationSTATS306B STATS306B. Clustering. Jonathan Taylor Department of Statistics Stanford University. June 3, 2010
STATS306B Jonathan Taylor Department of Statistics Stanford University June 3, 2010 Spring 2010 Outline K-means, K-medoids, EM algorithm choosing number of clusters: Gap test hierarchical clustering spectral
More informationThe Detection of Faces in Color Images: EE368 Project Report
The Detection of Faces in Color Images: EE368 Project Report Angela Chau, Ezinne Oji, Jeff Walters Dept. of Electrical Engineering Stanford University Stanford, CA 9435 angichau,ezinne,jwalt@stanford.edu
More informationRecognizing Handwritten Digits Using the LLE Algorithm with Back Propagation
Recognizing Handwritten Digits Using the LLE Algorithm with Back Propagation Lori Cillo, Attebury Honors Program Dr. Rajan Alex, Mentor West Texas A&M University Canyon, Texas 1 ABSTRACT. This work is
More informationImage segmentation. Václav Hlaváč. Czech Technical University in Prague
Image segmentation Václav Hlaváč Czech Technical University in Prague Center for Machine Perception (bridging groups of the) Czech Institute of Informatics, Robotics and Cybernetics and Faculty of Electrical
More informationClassification. Vladimir Curic. Centre for Image Analysis Swedish University of Agricultural Sciences Uppsala University
Classification Vladimir Curic Centre for Image Analysis Swedish University of Agricultural Sciences Uppsala University Outline An overview on classification Basics of classification How to choose appropriate
More information6. Object Identification L AK S H M O U. E D U
6. Object Identification L AK S H M AN @ O U. E D U Objects Information extracted from spatial grids often need to be associated with objects not just an individual pixel Group of pixels that form a real-world
More informationCOSC160: Detection and Classification. Jeremy Bolton, PhD Assistant Teaching Professor
COSC160: Detection and Classification Jeremy Bolton, PhD Assistant Teaching Professor Outline I. Problem I. Strategies II. Features for training III. Using spatial information? IV. Reducing dimensionality
More informationColocalization Algorithm. User s Guide
Colocalization Algorithm User s Guide Copyright 2008 Aperio Technologies, Inc. Part Number/Revision: MAN 0082, Revision A Date: March 7, 2008 This document applies to software versions Release 9.0 and
More informationEE 701 ROBOT VISION. Segmentation
EE 701 ROBOT VISION Regions and Image Segmentation Histogram-based Segmentation Automatic Thresholding K-means Clustering Spatial Coherence Merging and Splitting Graph Theoretic Segmentation Region Growing
More informationMR IMAGE SEGMENTATION
MR IMAGE SEGMENTATION Prepared by : Monil Shah What is Segmentation? Partitioning a region or regions of interest in images such that each region corresponds to one or more anatomic structures Classification
More informationSurface-based Analysis: Inter-subject Registration and Smoothing
Surface-based Analysis: Inter-subject Registration and Smoothing Outline Exploratory Spatial Analysis Coordinate Systems 3D (Volumetric) 2D (Surface-based) Inter-subject registration Volume-based Surface-based
More informationEpithelial rosette detection in microscopic images
Epithelial rosette detection in microscopic images Kun Liu,3, Sandra Ernst 2,3, Virginie Lecaudey 2,3 and Olaf Ronneberger,3 Department of Computer Science 2 Department of Developmental Biology 3 BIOSS
More informationCHAPTER 3 TUMOR DETECTION BASED ON NEURO-FUZZY TECHNIQUE
32 CHAPTER 3 TUMOR DETECTION BASED ON NEURO-FUZZY TECHNIQUE 3.1 INTRODUCTION In this chapter we present the real time implementation of an artificial neural network based on fuzzy segmentation process
More informationTwo-step Modified SOM for Parallel Calculation
Two-step Modified SOM for Parallel Calculation Two-step Modified SOM for Parallel Calculation Petr Gajdoš and Pavel Moravec Petr Gajdoš and Pavel Moravec Department of Computer Science, FEECS, VŠB Technical
More informationA GENETIC ALGORITHM FOR MOTION DETECTION
A GENETIC ALGORITHM FOR MOTION DETECTION Jarosław Mamica, Tomasz Walkowiak Institute of Engineering Cybernetics, Wrocław University of Technology ul. Janiszewskiego 11/17, 50-37 Wrocław, POLAND, Phone:
More informationMEASURING THE TOTAL BRAIN VOLUME USING IMAGE J
MEASURING THE TOTAL BRAIN VOLUME USING IMAGE J (1) Open Image J by double-clicking the desktop icon. An Image J menu bar will appear on the screen. (2) Open your first image file by selecting File, then
More informationNorbert Schuff VA Medical Center and UCSF
Norbert Schuff Medical Center and UCSF Norbert.schuff@ucsf.edu Medical Imaging Informatics N.Schuff Course # 170.03 Slide 1/67 Objective Learn the principle segmentation techniques Understand the role
More informationChapter 3 Image Registration. Chapter 3 Image Registration
Chapter 3 Image Registration Distributed Algorithms for Introduction (1) Definition: Image Registration Input: 2 images of the same scene but taken from different perspectives Goal: Identify transformation
More informationSlide07 Haykin Chapter 9: Self-Organizing Maps
Slide07 Haykin Chapter 9: Self-Organizing Maps CPSC 636-600 Instructor: Yoonsuck Choe Spring 2012 Introduction Self-organizing maps (SOM) is based on competitive learning, where output neurons compete
More informationClustering (Basic concepts and Algorithms) Entscheidungsunterstützungssysteme
Clustering (Basic concepts and Algorithms) Entscheidungsunterstützungssysteme Why do we need to find similarity? Similarity underlies many data science methods and solutions to business problems. Some
More informationInterpolating Silicon Photomultipliers
Interpolating Silicon Photomultipliers Peter Fischer, Heidelberg University, Germany (Presenter) Claudio Piemonte, FBK, Italy We present the novel Interpolating Silicon PhotoMultiplier (ISiPM) topology
More informationMODELING FOR RESIDUAL STRESS, SURFACE ROUGHNESS AND TOOL WEAR USING AN ADAPTIVE NEURO FUZZY INFERENCE SYSTEM
CHAPTER-7 MODELING FOR RESIDUAL STRESS, SURFACE ROUGHNESS AND TOOL WEAR USING AN ADAPTIVE NEURO FUZZY INFERENCE SYSTEM 7.1 Introduction To improve the overall efficiency of turning, it is necessary to
More informationUnsupervised Learning
Networks for Pattern Recognition, 2014 Networks for Single Linkage K-Means Soft DBSCAN PCA Networks for Kohonen Maps Linear Vector Quantization Networks for Problems/Approaches in Machine Learning Supervised
More informationSupplementary Materials for
advances.sciencemag.org/cgi/content/full/4/1/eaao7005/dc1 Supplementary Materials for Computational discovery of extremal microstructure families The PDF file includes: Desai Chen, Mélina Skouras, Bo Zhu,
More information