# Register now to access 7 million high quality study materials (What's Course Hero?) Course Hero is the premier provider of high quality online educational resources. With millions of study documents, online tutors, digital flashcards and free courseware, Course Hero is helping students learn more efficiently and effectively. Whether you're interested in exploring new subjects or mastering key topics for your next exam, Course Hero has the tools you need to achieve your goals.

6 Pages

### final1

Course: MJR 59, Fall 2009
School: Columbia
Rating:

Word Count: 1883

#### Document Preview

Object User-Defined Classifier based in a Neural Network with Optimal Selection of Visual Features using a Genetic Algorithm Manuel J. Reyes G. School of Engineering and Applied Science Columbia University, New York, NY USA INTRODUCTION The object recognition problem consist in: given some knowledge of how a certain objects may appear, plus an image of a scene possibly containing those objects, report which...

Register Now

#### Unformatted Document Excerpt

Coursehero >> New York >> Columbia >> MJR 59

Course Hero has millions of student submitted documents similar to the one
below including study guides, practice problems, reference materials, practice exams, textbook help and tutor support.

Course Hero has millions of student submitted documents similar to the one below including study guides, practice problems, reference materials, practice exams, textbook help and tutor support.
Object User-Defined Classifier based in a Neural Network with Optimal Selection of Visual Features using a Genetic Algorithm Manuel J. Reyes G. School of Engineering and Applied Science Columbia University, New York, NY USA INTRODUCTION The object recognition problem consist in: given some knowledge of how a certain objects may appear, plus an image of a scene possibly containing those objects, report which objects are present in the scene and where [1]. The rst step for the solution of this problem is to segment the image into regions that would presumably contain the desired objects. Some previous works on object recognition use the hypothesisveri cation paradigm. The generation of hypothesis is trivialized in the top-down case. If an object class has a stable property, they assign the object class as an interpretation for any image region that has that stable property; the main drawback of this approach is that the assignment has utility only in very restrictive scene domains, which subtract exibility to the systems. [1][2] In this project we implement a Neural Network as the object classi er-recognizer, the input of the neural network is a set of "useful" region features that would enable the neural network to classify a given region as a "object-container" or as a "nonobject-container" of the desire object. The set of "useful" region features is selected from a standard feature vector using a genetic algorithm. The genetic algorithm would eliminate the redundant and irrelevant features increasing the performance of the neural network both in speed and predictive accuracy. We present in this paper the results in the detection of faces in an image. The set of images chosen is highly unrestricted, in order to show that this approach is very exible. As we would see the proper selection of the features by the genetic algorithm increases the neural network accuracy in almost 20%. FACTORS TO CONSIDER IN THE SOLUTION'S DESIGN AND RELATED WORK. As we mentioned before, object recognition consist in classifying an image region as "container" or as "noncontainer" of a given object. The rst stage of an object 1 recognition system is the image segmentation module, which divides the image in homogeneous regions, which should correspond to the different objects in the image. The actual pixel data of the region cannot be use as the input of the neural network because the size of the segmented regions is not fixed, it varies in a large range and the running time of the neural network would be significant because the number of input units could be big. So we need to represent the information contained in the region in a more compacted and controlled way, we resolve this problem calculating a feature vector using the region pixel data to represent the characteristics of the given region. An optimal subset of this feature vector would be used as the input of the neural network. A previous work has used a neural network and a set of features from the image regions to recognize objects[3]. In this work they use a system of three layers of neural networks for the understanding of images. The classification of the objects is done using a set of classification rules (IF-THEN rules) that are encoded in the weights of a neural network using simple perceptron connections; this neural network corresponds to one of the mentioned layers, the applications of this system are also very constrained. Like in the previous mentioned work, in most of the object recognizer systems, the system designer using a set of classification rules predefines a model of the object. In our approach the user defines its own classifier training the neural network with a set of regions that the user has already classified as: "objectcontainer or as "non-objectcontainer", this "user definition" approach has been used in the work[4], but there, an algorithm using a variation of the K-Nearest-Neighbor Classification was used as the classifier, and the selection of the features used is not given by their "goodness" in the classification process. Feature selection plays a central role in the data analysis process since irrelevant features often degrade the performance of algorithms devoted to data characterization, rule extraction and construction of predictive models, both in speed and in predictive accuracy. Irrelevant and redundant features interfere with useful ones, so that most supervised learning algorithms fail to properly identify those features that are necessary to describe the target concept.[5] A genetic algorithm does the selection of the optimal subset; the number of features in this optimal subset would be also the input units number of the neural networks, so the feature selection has also a repercussion in the neural network architecture, in order to assure the best architecture for a specific subset of features, the genetic algorithm also encodes the number of hidden units to be used. So the genetic algorithm determines the optimal subset and the architecture of the neural network. The system could be trained to classify more than one object at the time, if this is the case the user should proportioned regions instances for each one of the objects to classify, this would change the number of outputs of the neural network 2 USE OF A NEURAL NETWORK OBJECT CLASSIFIER. AS AN As we mentioned before the image is segmented into regions with salient homogeneous properties, such as color and texture [6]. The parameters in the segmentation process should be calculated in such a way that the objects to be classified would be contained within a single region. Then a feature vector of all the regions present in the image is calculated. For this project 22 features were calculated: The first three correspond to the values of the region's average color in the LUV space, the next six features are related to the shape of the region, such as: roundness, aspect ratio, etc; the following four features are related to the size of the region: area, maximum perimeter, line and minimum line, the next six correspond to texture features of the region such as coarseness and contrast, finally the last three are related with the spatial location of the region with respect to the whole image. The user should determine which are the regions that contain the desired object, so all the regions in the image are classified as "object-container" or "no-object-container", this information is used as the target concept for the neural network to be implemented as the classifier. The information provided by the feature vector is used as the input of the neural network, but, as we mentioned before, because irrelevant and redundant features often interfere with useful ones, then the optimal subset is calculated and used as the input of the neural network. The complete feature vector contains 22 features, the first 19 are related to the intrinsic characteristics of the region, the last three give information about the location of the region within the image, so for classification purposes only the first 19 features would be used, the remaining 3 features are use only to report where a recognized object is located. USE OF A GENETIC ALGORITHM FOR OPTIMAL FEATURE SELECTION IN OBJECT CLASSIFIERS. The selection of the optimal subset of features is done using a genetic algorithm, the subset is encoded in a bit string, where a (1) in a certain position of the string denotes the selection of the feature in the given position and a 0 denotes the absence of the correspondent feature. The number of features selected would also determine the number of inputs of the neural network and then also would affect its architecture, so we also encoded in the bit string the number of hidden units of the neural network in the last three bits of the string. So the range of hidden units to be tested is from 1 (0000) to 16 (1111). An example of the bit used is the following: strings 1001100101000100000 0101 The first 19 bits determine which features are selected and the last three bits determine the number of hidden units of the neural network, for this particular case the number of hidden units encoded is six. The the genetic operators used are standard single-point 3 crossover [7]. and point mutation The selected set of features and the correspondent neural network architecture encoded in the bit string are used to train the neural network over a training set of segmented images with their correspondent regions previously classified by the user as container or noncontainer of the desired object. The fitness function used is the predictive accuracy of a given neural network over a test set of segmented images. The overall system can be seen in figure #1. Segmentation Module (S.M.) Image Segmented Image Feature Extraction Module S.M. Training Set S.M. User-Classified Regions Input Test Set Feature Extraction Module Neural Network Output Control Signals Genetic Algorithm Accuracy measurement in the test set Training phase Fitness Function Figure # 1 4 RESULTS The system was used to construct a face detector. We use a training set of 200 images containing faces in different orientations of persons of different races and ages; and a test set containing 400 images with the same diversity as in the training set. This is a very unconstrained set, because normally in face detectors the image of the face is straight or is constrained in terms of the race or age. The accuracy obtained using the entire feature vector using 16 hidden units was 70.45%. The value of the parameters used in the neural network and in the genetic algorithm are the follow...

Find millions of documents on Course Hero - Study Guides, Lecture Notes, Reference Materials, Practice Exams and more. Course Hero has millions of course specific materials providing students with the best way to expand their education.

Below is a small sample set of documents:

Johns Hopkins - IBM - 7094
Johns Hopkins - OS - 3
Johns Hopkins - OS - 3
Columbia - BC - 2003
Organic Contaminants' Affinity for Protozoan-Modified Dissolved Organic Matter (DOM)Nella Green, Environmental Science Dep't, Barnard Elizabeth Kujawinski, Research mentor, Barnard Dallas Abbott, Seminar advisor, LDEOProtozoa: Flagellates &amp; Ciliat
Columbia - CS - 4162
To Do / MotivationAdvanced Computer Graphics (Spring 2006)COMS 4162, Lecture 13: NURBs, Spline Surfaces Ravi Ramamoorthihttp:/www.cs.columbia.edu/~cs4162Questions? One of written questions in assignment deals with material in this lecture Otherw
Columbia - TL - 2383
Arithmetical Denability over Finite StructuresTroy Lee CWI and University of Amsterdam December 17, 2003Abstract Arithmetical denability has been extensively studied over the natural numbers. In this paper, we take up the study of arithmetical dena
Columbia - MB - 95
Columbia - MB - 84
Columbia - MB - 80
Columbia - SRR - 2001
RESEARCH IN REGIONAL SEISMIC MONITORING F. Ringdal1, E. Kremenetskaya 2 , V. Asming2, T. Kvrna1 , S. Mykkeltveit1 , J.I. Faleide1 and J. Schweitzer11 NORSAR 2KolaRegional Seismological CenterSponsored by Air Force Technical Applications Center C
Columbia - SRR - 2000
EVALUATION OF INFRASONIC SPATIAL FILTERS MICHAEL A.H. HEDLIN, JON BERGER AND MARK ZUMBERGE INSTITUTE OF GEOPHYSICS AND PLANETARY PHYSICS UNIVERSITY OF CALIFORNIA, SAN DIEGO SPONSORED BY DEFENSE THREAT REDUCTION AGENCY ABSTRACT The infrasound element
Columbia - SRR - 2000
PHENOMENOLOGY RESEARCH USING PAST NEVADA TEST SITE EXPLOSION AND EARTHQUAKE DATA Kevin M. Mayeda, Michael E. Pasyanos, Jennifer OBoyle Stephen C. Myers, William R. Walter, and Peter Goldstein Lawrence Livermore National Laboratory Sponsored by the U.
Columbia - SRR - 2000
EXCITATION AND ATTENUATION OF REGIONAL WAVES, AND MAGNITUDE DEPENDENCE OF PN/LG RATIOS IN EASTERN EURASIA Jiakang Xie, Lamont-Doherty Earth Observatory of Columbia University Sponsored by The Defense Threat Reduction Agency Contract No. DSWA01-98-1-0
Columbia - SRR - 2002
24th Seismic Research Review Nuclear Explosion Monitoring: Innovation and IntegrationGROUND TRUTH COLLECTION FOR MINING EXPLOSIONS IN NORTHERN FENNOSCANDIA AND RUSSIA David B. Harris,1 Frode Ringdal,2 Elena Kremenetskaya,3 Svein Mykkeltveit,2 Donal
Columbia - WTD - 1
44 1 Ch21: Confucianism in the Early Edo Period Confucianism had originally been imported along with the introduction of Chinese culture sponsored by Prince Shtoku and the Chinese administrative and political model that was implemented through the so
Columbia - WTD - 2002
MAJOR TOPICS IN EAST ASIAN CIVILIZATIONS Session #48 Korea 1850-1998 I. Korea in the late nineteenth century A. Challenges fro m within 1. Financial insolvency 2. Corruption 3. Social unrest 4. Decline of the royal house B. Challenges from without 1.
Columbia - WTD - 2002
MAJOR TOPICS IN EAST ASIAN CIVILIZATION Session #1Introduction to East Asian Civilization I. Defining Features of East Asian Civilizations A. Geographical Scope: The common higher civilization of the East Asian area occupied by the modern countries
Columbia - WTD - 2002
MAJOR TOPICS IN EAST ASIAN CIVILIZATIONS Session # 37 Japanese Responses to the W estI.Japan in the early nineteenth century A. B. Social change, economic crisis, and intellectual disquiet The condition of the shogunate, and the challenge of outl
Columbia - WTD - 2002
MAJOR TOPICS IN EAST ASIAN CIVILIZATION Session #7 Leg alism and the Q in (Ch'in) D ynasty I. Legalist &quot;rea lism &quot;: Its ins titution al form s (law s, punishm ent, system s, etc.). The Legalist ideal of universal law; its impersonality and autom atic
Columbia - WTD - 2002
M A JOR TOPICS IN EA ST A SIAN CIVILIZATION Session s # 4 0Liberal D em oc rac y , So cialism and Nat ion alism in 2 0 t h c ent ury East A sia, Part I Jap an I. Liberal p olit ics an d so cial c rit icism in t he M eiji and Taish o p eriod s (1 9
Columbia - WTD - 2002
MAJOR TOPICS IN EAST ASIAN CIVILIZATIONS Session #35 China 1644-1840 I. Qing state and so ciety A. B. C. Th e co nqu est of Ch ina; the ban ner system of ga rrison state The reestablishment of major Chinese civil institutions Zhu Xi Neo-C onfucianism
Columbia - WTD - 2002
MAJOR TOPICS IN EAST ASIAN CIVILIZATION Session #5 Confucianism Confucius: early texts preserved in the Confucian Canon: Historical Documents, Odes, Rites, Changes. The ideals of the Confucian sage-king, Heaven, the &quot;Mandate&quot; and the revival of the Z
Columbia - WTD - 2002
MAJOR TOPICS IN EAST ASIAN CIVILIZATION Session #9 Han Confucianism Dong Zhongshu (Tung Chung-shu) and the Confucian adjustment to irreversible changes in the political situation: accomm odation to bureaucratic state; timeless values and action in ti
Columbia - WTD - 2002
MAJOR TOPICS IN EAST ASIAN CIVILIZATIONS Session # 18 State Building in the Late Yamato and Nara Periods, 645 CE-794 CEI.Taika Reforms, 645 CE A. Adoption of a Chinese-style bureaucratic state system B. Reorganization of the landholding systemI
Columbia - WTD - 1
16 1 Ch. 24: Confucian Revisionists Fundamentalism and Revisionism in the Critique of Neo-Confucianism Neo-Confucianism in seventeenth- and eighteenth-century Japan, as represented by Zhu Xi's synthesis of speculative thought in the Song school, was
Columbia - WTD - 1
52 1 Ch. 30: Eighteenth-Century Rationalism As Japan moved into the eighteenth century the Tokugawa Shogunate was already a century old, with its power and prestige firmly established and its policy of peace enforcement a proven success. The country
Columbia - WTD - 1
1Ch. 36: The Debate Over Seclusion and Restoration After 1739 Russian ships were seen in Japanese waters with increasing frequency. Areport was brought home by waifs that the Russians had established a school of navigation at Irkutsk in 1764, and
Columbia - WTD - 1
1Ch. 32: The National Learning Schools The earliest extant writings in Japan are about Japan, and so it is no exaggeration to saythat Japanese interest in Japan-its mythology, heritage, culture and traditions-is as old as Japanese history itself.
Columbia - WTD - 1
1Ch. 31: The Way of the Warrior II The Way or ethos of the Warrior, which achieved its first notable articulation in medievaltimes (Chapter 12), continued to assert itself under the Tokugawa shogunate, a basically military regime. As we have seen
Columbia - WTD - 1
1Ch. 25: Varieties of Neo-Confucian Education Whatever the different views espoused by individual thinkers of the Edo period, asshown above, Neo-Confucian education had a significant life of its own. As in China and Korea during the same period,
Columbia - WTD - 1
1Ch. 22: The Spread of Neo-Confucianism in Japan The prestige of Neo-Confucianism as an officially approved teaching arose in part fromthe support given it by leading members of the Tokugawa family. Among the many sons of Ieyasu who contributed t
Columbia - WTD - 1
Ch 21: Confucianism in the Early Edo Period Confucianism had originally been imported along with the introduction of Chinese culture sponsored by Prince Shtoku and the Chinese administrative and political model that was implemented through the so-cal
Columbia - WTD - 2002
MAJOR TOPICS IN EAST ASIAN CIVILIZATION Session #29 &amp; 30 Ming Dynasty China (1368-1644) Ming Dynasty, the Restitution of Chinese Rule, 1368-1644 Capitals: Nanking and Peking I. Found ing of the M ing D ynasty -Zhu Yuanzhang as emperor: reign of Hongw
Columbia - WTD - 1
Ch. 22: The Spread of Neo-Confucianism in Japan The prestige of Neo-Confucianism as an officially approved teaching arose in part from the support given it by leading members of the Tokugawa family. Among the many sons of Ieyasu who contributed to it
Columbia - RJS - 19
Identity crisis: The literary cult and culture of Odessa in the early twentie.Rebecca Stanton Symposium; Fall 2003; 57, 3; ProQuest Direct Complete pg. 117Reproduced with permission of the copyright owner. Further reproduction prohibited without p
Columbia - DBL - 11
David Barnett LurieAssistant Professor of Japanese History and Literature Department of East Asian Languages and Cultures Columbia University39 Claremont Avenue, Apt. 53 New York, NY 10027 (212) 663-6499 DBL11@columbia.eduEDUCATIONc/o EALAC, 40
Columbia - WW - 2040
Service Engineering 2002September 2,Homework 7: GazolCo's Call Center*Ten agents are busy answering calls at GazolCo's call center. Most calls are by customers calling to pay or inquire about their gas bills. Looking through recent ACD reports y
Columbia - PATH - 2
Pathways of Bilirubin Synthesis and CatabolismBone MarrowHgb Globin + Heme Fe + PorphogensRBCsReticuloendothelial System (RES)Hgb Biliverdin Bilirubin Fe Globin Heme Precursors Myoglobin Non-Hgb Heme Proteins Shunt PathwayFree Bilirubin Alb
Columbia - PS - 1
Clinical Practice Lecture 1/22/03 Dr. Ora Pearlstein Transcribed by Perry Wilson &quot;Boundaries&quot; -Take Home Message: It's never ethical to have sex with your patients or former patients if you're a psychiatrist. It might be OK to have sex with former (n
Columbia - PATH - 2
Hemolytic anemia 2/12/04 Hemolytic anemias are those that result from an increase in the rate of red cell destruction. The hallmark of these anemias is shortened red cell survival; the red cells are typically normochromic and normocytic. Other charac
Columbia - PATH - 2
Hematology 6: Platelet Function Dr. Druiguid (24 February 2003) See also last years transcript, which is almost identically the same info but in a more narrative form, if thats your kind of thing . . . Circulatory System: The circulatory system has e
Columbia - PATH - 1
M&amp;CP Lecture 4 8/28/03 Lecturer: Dr. Buttyan Transcriber: Dara Liotta APOPTOSIS Cell death occurs under both normal and abnormal conditions Development of a whole human being from a single fertilized ovum requires generation of about a trillion cells
Columbia - PATH - 1
Lecturer: Paul Rothman (pbr3@columbia.edu) Transcriber: David Parris (dlp2006@columbia.edu) Immunology lecture 3:Chemistry, Genetics, and Structure of Antigen Receptors IThis will be organized by slides so you can follow along when they end up sur
Columbia - PATH - 1
Immunology 4 Introduction to Microbiology Drs. Shuman and Young 8/29/03 Transcriber: Dan Hwang Introduction to Bacteriology Bacteria -Size ~10-6 m (around 1000 will fit into an animal cell) -Focus on important structures of bacteria and how they ar
Columbia - PATH - 1
MID/Micro 4 10/2/02, 10:00 am lecture Dr. Figurski Beth Harre (esh16) Note: Waiting for edits from Dr. Figurski. GENETICS III Conjugation is the most prevalent from of gene transfer in bacteria, and can be very efficient. It requires that 2 bacterial
Columbia - PATH - 2
Columbia - PATH - 1
Shock/ Hemorrhage/ Thrombosis Shock- A low-perfusion circulatory insufficiency state leading to an imbalance between metabolic needs of vital organs and available blood flow. -The tissues are effected by: 1. decreased oxygen and nutrient delivery 2.
Columbia - PATH - 1
MID Lecture #6 Dr. Shuman: Bacterial Physiology 10/11/02 Chris Daley (ctd2003) Major theme: Bacteria have a lot of physiological things in common. There are also areas with a lot of diversity. When thinking about the physiological differences, think
Columbia - PATH - 2
Hematology Transcript: Acute Leukemias Dr. David Savage, lecturer Noah Raizman, transcriber 1. Acute vs. Chronic Leukemias: -Acute Leukemias are characterized by monoclonal proliferation of hematopoietic precursor cells, or blasts, which have the fol
Columbia - PATH - 1
Dana Critchell dc2016@columbia.edu Dr. Despommier Lecture #3 (10/21/02) Cestodes Dr. Despommier started out by going over how to treat the parasites that we discussed last week. -Treat hookworms with Mebendazole -Treat Strongyloides stercoralis with
Columbia - PATH - 1
MID - 45 Rickettsia, Ehrlichia, and BorreliaLecturer: Dr. Deborah Rudin, M.D. saverp@aol.com Transcriber: Marc Russo mwr2002@columbia.eduWho's at risk? those exposed to areas where they can be bitten by a tick/louse/mite or work with animals: farme
Columbia - PS - 4
Teaching To Teach Seminars for Medical Students January 2006Module 1: Assessing The LearnerTeaching To Teach Faculty Pablo Joo, MD Director of Predoctoral Education, Center for Family Medicine Mark Graham, Ph.D. Director of Education Researc
Columbia - PS - 4
Module 2: Feedback Skills Teaching To Teach SeminarsPrepared by the Society of Teachers of Family Medicine Presented by Pablo Joo MD Center for Family Medicine / NYP / P&amp;STeaching To Teach Sponsors Center for Education Research and Evaluation Ce
Columbia - PATH - 1
5- Protozoa: Malaria (Plasmodium falciparum, Plasmodium vivax, Plasmodium ovale, Plasmodium malariae)General Protozoa/Malaria Info: Single celled eukaryotic organisms Equatorial disease, worldwide, very prevalent in Africa Most are free-living L
Columbia - PATH - 1
6- AIDS Opportunistic Infections (Toxoplamsa gondii &amp; Pneumocystis carinii/jiroveci)Protozoan Toxoplasm a gondiipg 68HostsCats, humansLife CycleDEFINITIVE HOST (CATS): 1) cats ingest oocysts in cat feces or pseudocysts in infected carrion Ing
Columbia - PATH - 1
Enterobius vermicularis* Embryonated eggsEnterobius vermicularis Adult FemaleTricuris trichiura Adult female and adult maleEnterobius vermicularis* (pinworm) Embryonated eggs (see larvae inside) Symptoms- itchy perianal region Diagnose- eggs in
Columbia - PATH - 1
7-Protozoa that cause diarrheal disease: Giardia lamblia, Entamoeba histolytica, Cryptosporidium parvumProtozoan Host Life Cycle, Pathology Beavers, 2 stages: Giardia dogs, humans trophozoite lambliaPg 7Clinical Disease, DiagnosisCLINICAL PRESEN
Columbia - PS - 1
Columbia - PS - 1
Columbia - PS - 1