Sample ScriptFaculty of Engineering and EnvironmentResearch proposalModule: KF7028 Research and Project ManagementModule tutor: Shelagh Keogh/Rebecca Strachan/Perla InnocentiAssignment title: MSc Research ProposalStudent Name: (note direct ane indirect personalinfomrmatio has been removed)Students Number:Programme of study title: MSc computer scienceSupervisor’s name: Ammar BelatrecheSecond Marker’s name: Shelagh KeoghAcademic Year:Viva Date: 27/03/201 Sample Script1Asmail Muftah 14042917Contents1 Aim……………………………………………………………………………………………………………………………………. 32 Background, Motivation and Relevance – literature review ……………………………………………………… 32.1 Background ………………………………………………………………………………………………………………….. 32.2 Motivation …………………………………………………………………………………………………………………… 42.3 Literature Review………………………………………………………………………………………………………….. 52.3.1. Unsupervised methods……………………………………………………………………………………………. 5Supervised methods ………………………………………………………………………………………………………… 52.4 Sources and use of Knowledge ……………………………………………………………………………………….. 62.4.1 Journal …………………………………………………………………………………………………………………… 62.4.2 The aspects of standards for publication in IEEE:………………………………………………………… 62.4.3 Relevant Authors and Journals …………………………………………………………………………………. 73 Scope, Objectives and Risk …………………………………………………………………………………………………… 83.1 Scope…………………………………………………………………………………………………………………………… 83.2 Objective……………………………………………………………………………………………………………………… 83.3 Risk……………………………………………………………………………………………………………………………. 114 Ethics, Legal, Social, Security and Professional Issues …………………………………………………………….. 124.1 Ethics…………………………………………………………………………………………………………………………. 124.2 Legal………………………………………………………………………………………………………………………….. 124.3 Social…………………………………………………………………………………………………………………………. 124.4 Security ……………………………………………………………………………………………………………………… 124.5 Professional………………………………………………………………………………………………………………… 125 Schedule of Activities…………………………………………………………………………………………………………. 125.1 Work Breakdown Structure (WBS) ………………………………………………………………………………… 135.2 Tasks list…………………………………………………………………………………………………………………….. 145.3 Gantt chart…………………………………………………………………………………………………………………. 156 References ……………………………………………………………………………………………………………………….. 16STUDENT PROJECT: ETHICS REGISTRATION AND APPROVAL FORM ……………………………………… 182Asmail Muftah 14042917List of FigureFigure 1 Mind map ………………………………………………………………………………………………………………….7Figure 2 Proposed blood vessel segmentation system …………………………………………………………………9Figure 3 Processes of pre-processing and feature extraction …………………………………………………….. 10Figure 4 Work breakdown Structure ………………………………………………………………………………………. 13Figure 5 Gantt chart ………………………………………………………………………………………………………………15List of TablesTable 1 Relevant Authors and Journals ………………………………………………………………………………………8Table 2 Risk form………………………………………………………………………………………………………………….. 11Table 3 Tasks list ………………………………………………………………………………………………………………….14Sample Script3Asmail Muftah 14042917Blood vessel segmentation in retinal images1 AimBlood vessel segmentation in retinal images is an important stage in retinalimage analysis and computer-aided diagnosis of various ophthalmologic diseases.The aim of this project is to develop an effective and efficient technique for bloodvessel segmentation in retinal images.2 Background, Motivation and Relevance – literaturereview2.1 BackgroundSeveral methods for retinal image segmentation have been reported in theliterature. Based on machine learning methods, the segmentation of blood vessels inretinal images could be classified into 2 groups: supervised methods (Marin et al.,2011; Ricci and, 2007; Perfetti Soares et al., 2006) and unsupervised methods (Kandeet al., 2010; Rahebi and Hardalaç, 2014; Villalobos-Castaldi et al.,2010). Supervisedmethods are based on the prior pixel labelling information which classifies a pixel as avessel or non-vessel (i.e. background). Whereas, unsupervised methods do not useprevious labelling data and they can learn and arrange data to discover the patternsor clusters that assimilate the blood vessels. Filtering or kernel-based methods usea Gaussian shaped curve to model the blood vessel cross-section and rotate thematched filters to detect blood vessels with different orientations (Zolfagharnasab andNaghsh-Nilchi, 2014). Different shaped Gaussian filters such as simple Gaussianmodel (Zolfagharnasab and Naghsh-Nilchi, 2014) and derivative of Gaussian function(Zhang et al., 2010) have been used for blood vessel detection. Another method basedon mathematical morphology (Fraz et al., 2012), takes advantage of known vesselfeatures and boundaries and represents them in mathematical sets. Then, utilisingmorphological operators, the blood vessels will be extracted from the background. Byfollowing blood vessel centre lines, the vessel tracking based methods as proposedby Fraz et al. (2012) try to acquire the vasculature structure (Fraz et al., 2012). Usually,a set of starting points is established and then the vessels’ traces are generated based4Asmail Muftah 14042917on local information, attempting to discover the way that best matches the blood vesselprofile model. In model-based techniques, clearly stated vessel models are applied todetect the blood vessels (Al-Diri, Hunter and Steel, 2009).2.2 MotivationManual segmentation of the blood vessels from retinal images is exhausting,tedious and time-consuming. In addition, if the vascular network is highly complicated,making detailed segmentation becomes challenging (You et al., 2011). Therefore,automated segmentation is significant, as it saves both time and effort required formanual segmentation, and in an ideal situation of blood vessel segmentation, theresults of segmentation that come from an automated algorithm can be better than theresults of specialist manual labelling (Wasan et al., 1995). To reach practicalapplications, the algorithms have to be easy to use for the user who is non-specialistin this technology, in this case, algorithms should not fundamentally rely on upondesigning various parameters (Sarah and Shadgar, 2009). There are challengesfacing automated blood vessel segmentation such as low contrast in images, anextensive variety of vessel widths and an assortment of various structures in retinalimages e.g. retinal image boundaries, optic disc and retinal lesions caused bydiseases (Mendonca and Campilho, 2006). Even though different techniques areavailable for segmentation of blood vessels from retinal images, there is still a needfor improving the segmentation accuracy and efficiency.Mostly, blood vessel segmentation algorithms concentrate on automatic detectionrelated to diabetic retinopathy, which in recent days is considered as the major causeof blindness. Vision loss related to diabetic retinopathy can be prevented if the diseaseis discovered at an early stage (Taylor and Keeffe, 2001).Based on different algorithms, a diverse range of blood vessel segmentationapproaches have been proposed (Sum and Choung, 2008; Kuppusamy and Divya,2016). There is a difference among the algorithms in terms of complexity andsegmentation performance. Different algorithms for blood vessel segmentation will bereviewed in this project in order to identify the most effective and efficient techniquesfor blood vessel segmentation in retinal images. This project will evaluate existingtechniques of blood vessel segmentation and develop this technique, to produce moreeffective and efficient algorithms.Sample Script5Asmail Muftah 140429172.3 Literature ReviewThe segmentation of blood vessel in retinal images could be divided two classes:supervised methods and unsupervised methods.2.3.1. Unsupervised methodsUnsupervised methods of blood vessel segmentation can be grouped into few subclasses: matched filtering (Chaudhuri et al., 1989), vessel tracking (Liu and Sun,1993), morphological processing (Zana and Klein, 2001; Mendonca and Campilho,2006) and model based (Delibasis et al., 2010). Chaudhuri et al. (1989) first proposedmatched filtering method of the blood vessels segmentation in retinal images. Inhealthy images the matched filtering work well however it raise the false positives. Ablood vessel tracking method was utilised in retinal images (Liu and Sun, 1993) byfollowing local vessel with selecting a set of vessel initial by using a given vessel initialas starting points. To extract large vessels, these methods are efficient, but withoutseed points usually it fails. Zana and Klein (2001) combined Morphological filters andcurvature to detect local linear and connected vessel structure. The detection of largevessels was good however the thin vessels were missing. Mendonca and Campilho(2006) presented a method to detect the blood vessel tree by pixel processingBased methods on tracking methods for vessel centre lines detection were usedin the retinal angiogram, the dataset was DRIVE database and STARE database, theAverage Accuracy for DRIVE images was (94.52%), whereas, the Average Accuracyfor STARE images was (94.77%). To extract retinal blood vessels Zhao et al. (2014)exploited the region growing technique and a district based active contour model withlevel set implementation. The results of this method reached the accuracy of 95.05%on the STARE images and 94.77% on the DRIVE images. Besides, the multi-scalemethod was implemented to segment vessels due to the different widths and directionof the retinal blood vessels. This technique exploits the geometric model without userintervention (Delibasis et al., 2010).Supervised methodsIn supervised methods, each pixel is classified as a vessel or a non-vessel (i.e.,background) whereas set features represent a pixel. The classifier uses vectors offeature extraction from the labelling information to distinguish vessel from non-vesselpixels. Ricci and Perfetti (2007) used support vector machine classification and line6Asmail Muftah 14042917operators to segment blood vessels in retinal images. A DRIVE database of colourfundus images used, because they avoided pre-processing and low pass filtering thethinnest vessels have been lost, and the accuracy was 91.47 %. A 41-D feature vectorand AdaBoost classifier were proposed by Lupascu, Tegolo and Trucco, (2010) wherethe features acquired from multiple scale matched filters. The algorithm tested on theDRIVE dataset and achieved an accuracy of 95.97 %. Marin et al (2011) used a multilayer feed-forward neural: network and a: 7-D feature vector, which is: composed of:gray-level and: moment invariant-based: features. This method used the DRIVE andthe STARE datasets and achieved better accuracy on the STARE dataset. Theaccuracy using the STARE dataset was 0.9526, while the accuracy on the DRIVEdataset was 0.9452. Staal et al (2004) used a k-NN classifier based on image ridgesextraction, this method tested on a DRIVE database of 40 images, which labelledmanually. The accuracy of this method is 94.70 %. Soares et al. (2006) used asGaussian mixture a Bayesian classifier: with class-conditional: probability: densityfunctions on DRIVE database and STARE Database the result were 94.66 % for DRIVEdatabase and 94.80 % for STARE database.2.4 Sources and use of Knowledge2.4.1.JournalThere are several journals relevant to the topic of blood vessel segmentation in retinalimages. IEEE Transactions published the most relevant journal papers to this topicon Medical Imaging. In additional, most of the academics consider IEEETransactions on Medical Imaging as a good journal, and it has many publications onBlood vessel segmentation in retinal images where these papers use differenttechniques and have been published in last decade.2.4.2.The aspects of standards for publication in IEEE:2.4.2.1 Page LayoutThe page size must be A4, the margins in the top = 19 mm, in the bottom =43mm and left = right = 14,32mm2.4.2.2 Page StyleAll paragraphs in the paper must be indented and justified.Times New Roman or Times font should use the text front of the entiredocument. The Regular font of the Title must be 24 pt and 11 pt for the AuthorSample Script7Asmail Muftah 14042917name where 10 pt should use for Author affiliation. Paragraphs and level 1, 2and 3 must be in 10 pt Regular font. Abstract body, author email address andcell in a table must be in 9 pt table caption, figure caption and reference itemmust be in 8 pt the number of the reference use for referring to a reference.2.4.2.3 Reference formatThere are a different format for the reference, and these are the most importanttwo.Book format“A. A. Author, Title: Subtitle (in italics), Edition (if not the first), Vol.(if amultivolume work). Place of publication: Publisher, Year, page number(s) (ifappropriate)” (Anon, 2017). Journal format“A. A. Author of article. “Title of article,” Title of Journal, vol. #, no. #, pp. pagenumber/s, Month year” (Anon, 2017).2.4.3.Relevant Authors and Journals Title ofJournalsTitle of articleAuthorsYearImpactfactorPatternrecognitionAn effective retinal blood vesselsegmentation method usingmulti-scale line detectionNguyen, U. T., Bhuiyan, A.,Park, L. A., &Ramamohanarao, K20133.399 Mind MapRetinalvesselsegmenRetinalimageMedicalimageprocessingImagesegmentatiBloodvesselsegmentFigure 1 Mind map8Asmail Muftah 14042917 ExpertSystems withApplicationsAn unsupervised coarse-to-finealgorithm for blood vesselsegmentation in fundus imagesNeto, L.C., Ramalho, G.L.,Neto, J.F.R., Veras, R.M.and Medeiros, F.N20172.981IEEEtransactions onmedicalimagingA cross-modality learningapproach for vesselsegmentation in retinal images.Li, Q., Feng, B., Xie, L.,Liang, P., Zhang, H. andWang, T20163.756 Table 1 Relevant Authors and JournalsWe have chosen the journal of IEEE transactions on medical imaging because it hasa good impact factor. In additional, one of its focus is on image presses in the scienceof medicine.3 Scope, Objectives and Risk3.1 ScopeThe scope of this project is blood vessel segmentation in rental images. Thesegmentation method to develop should be as accurate and reliable as possible. Themain aim of segmentation is to differentiate an object of interest and the backgroundfrom an image existing blood vessel segmentation techniques will be evaluated anddevelop a more effective and efficient approach.The project is required to include these scopes:1- To explore tools, techniques and existing systems for segmenting blood vesselsin the literature2- To analyse and develop a more effective and efficient algorithm for thesegmentation of blood vessels in retinal images.3- To analyse and evaluate the proposed system in comparison with existing work.The diagnosis and checking of eye diseases such as glaucoma, diabetic retinopathyand hypertension will not be included.3.2 Objective1- Critical Review of the Literature1.1 Explore and critically review the existing literature of blood vessel segmentation inretinal images.Sample Script9Asmail Muftah 140429171.2 To obtain the efficient method that used in segment blood vessel from retinalimages.1.3 Study the Structure of the human eye.To have knowledge about human eye which will help in blood vesselsegmentation.1 Image processing techniques.To remove any unwanted features in the images such as noisy features that couldaffect the vessel segmentation system.1.3 Image enhancement and pre-processing.1.4 Detection of Blood VesselsIn order to segment blood vessel from retinal images.1.5 Vessel segmentationFigure 2 Proposed blood vessel segmentation system2 Features extraction10Asmail Muftah 14042917Before the process of classification, the image features have to be extractedwhich will help to find the most relevant features to segment the blood vesselfrom retinal images.Figure 3 Processes of pre-processing and feature extraction 3ClassificationThere are supervised method for image classification; this dissertation will use a few classification techniques. To find out the best techniques used inblood vessel segmentation. 4.1 Evaluate Artificial Neural Network4.2 Evaluate t Decision Tree4.3 Evaluate Support Vector Machine (SVM)4.4 Evaluate K-Nearest Neighbour (KNN)4 Post processingThe post-processing stage is another important operation to obtain better andaccurate segmentation. The output of blood vessel segmentation has to pass througha module of post processing which will fill the pixel gaps on the detected of vesselsand removed the detected of non-vessels.Sample Script11Asmail Muftah 140429173.3 RiskTable 2 Risk form Risk EventRiskValue(1-100)RiskMonitoring/Control FlagRiskManagementStrategyRisk ReviewdateTThe topic is new forme, and myKnowledge aboutthe topic is few2550delay duringthe researchRead as muchas possible onthe topicDuringresearchTIf there any need towork on weekenddays the lab will notbe open2550When I wentto the lap atweekend, Ifound itclosed1 Mange thetime to beworking days inthis project as itin work plan 2install thesoftware onlaptopDuringresearchTImplement the dataon ML it is the firsttime I will work ondataset of images,any error during theimplementationCould causing delay3440ErrorAsk thesupervisor forhelpDuringresearchFThere is no risk onpart of financial000////PThere is no risk forthe people000////ENo risk on the part ofEnvironment000////SSecurity have no risk00//// 12Asmail Muftah 140429174 Ethics, Legal, Social, Security and ProfessionalIssues4.1 EthicsIn terms of ethics, this project will not carry out any risks because the dataset that will be used in this project already provided for other researchers in asimilar field. There are no issues regarding to data confidentiality because thedata is already in the public domain.4.2 LegalThe project will not carry out any legal issues as the database already in thepublic domain. In addition, as defined by the law this research have not anywork with participants as well as any vulnerable groups.4.3 SocialIn terms of social, this project will not carry out any problem for the society.The project will be pleased in a simulation and will not engage human resourcesor organisations. Furthermore, this research will not work any participants orany vulnerable groups.4.4 SecurityAs indicated above the project will use dataset, which is already in the publicdomain and will not work with people, wherefore, the project will not carry outany security issues.4.5 ProfessionalThe project has no professional issues because the project will not work withpeople. Moreover, the dataset that will be used in this research is alreadypublished for the public.Finally, the ethics form of Northumbria Ethics Registration has been completed andexecuted with this research in order to be followed during the project period.The attachment of this form is in Appendix 1.5 Schedule of ActivitiesSuccessful projects rely on suitable planning and scheduling the tasks in the rightorder. Therefore, the Dissertation project plan had divided into a few objectivesSample Script13Asmail Muftah 14042917and tasks to guarantee the execution of the whole work. Table 1 represents thestructure of the task list. The work breakdown structure of this project is illustratedin Figure 3, while Figure 4 represents the Gantt chart, which is created to hold theresearcher on track. The Project will start on 22nd May 2017 and end on 14thSeptember 2017. During the semester, the working days are 84 days. The projectplan excluded the weekend days. However, the weekends days will be included ifthere is any need.5.1 Work Breakdown Structure (WBS)Figure 4 Work breakdown Structure14 | P a g e5.2 Tasks list TaskNo.Task nameDurationStartEnd1Literature Review10 Days22/05/201704/06/20171.1Carry out which method used in the literature4 Days22/05/201725/05/20171.2Study the Structure of the human eye2 Days26/05/201729/05/20171.3Criticise the literature review4 Days30/05/201702/06/20172Image processing techniques15 Days05/06/201725/06/20172.1Image enhancement and pre-processing5 Days05/06/201709/06/20172.2Detection of Blood Vessels5 Days12/06/201716/06/20172.3Vessel segmentation5 Days19/06/201723/06/20173Features extraction4 Days26/06/201729/06/20174Classification16 Days30/06/201723/07/20174.1Evaluate Artificial Neural Network4 Days30/06/201705/07/20174.2Evaluate Decision Tree4 Days06/07/201711/07/20174.3Evaluate Support Vector Machine4 Days12/07/201717/07/20174.4Evaluate K-Nearest Neighbor (KNN)4 Days18/07/201721/07/20175Post processing5 Days24/07/201730/07/20176Research draft writing33 Days31/07/201714/09/20176.1Writing Introduction and literature reviews6 Days31/07/201707/08/20176.2Writing Methodology5 Days08/08/201714/08/20176.3Writing Result and Discussion5 Days15/08/201721/08/20176.4Writing Conclusion, Recommendations5 Days22/08/201728/08/20176.5Proofreading and reviewing the dissertation5 Days29/08/201704/09/20176.6Reviewing the thesis and making the poster7 Days05/09/201713/09/20177Submission date0 day14/09/201714/09/2017meeting with the supervisor22nd May201714th Sep.201745 mins/week Sample Script1514Table 3 Tasks list5.3 Gantt chartFigure 5 Gantt chart166 ReferencesAl-Diri, B., Hunter, A. and Steel, D., 2009. An active contour model for segmenting andmeasuring retinal vessels. IEEE Transactions on Medical imaging, 28(9), pp.1488-1497.Anon, (2017). [online] Available at: http://paginas.fe.up.pt/~jca/wrsc/templates/IEEEConferenceA4-format.pdf [Accessed 22 Mar. 2017].Chaudhuri, S., Chatterjee, S., Katz, N., Nelson, M. and Goldbaum, M., 1989. Detection of bloodvessels in retinal images using two-dimensional matched filters. IEEE Transactions on medicalimaging, 8(3), pp.263-269.Delibasis, K.K., Kechriniotis, A.I., Tsonos, C. and Assimakis, N., 2010. Automatic model-basedtracing algorithm for vessel segmentation and diameter estimation. Computer methods andprograms in biomedicine, 100(2), pp.108-122.Fraz, M.M., Remagnino, P., Hoppe, A., Uyyanonvara, B., Rudnicka, A.R., Owen, C.G. andBarman, S.A., 2012. Blood vessel segmentation methodologies in retinal images–a survey.Computer methods and programs in biomedicine, 108(1), pp.407-433.Kande, G.B., Subbaiah, P.V. and Savithri, T.S., 2010. Unsupervised fuzzy based vesselsegmentation in pathological digital fundus images. Journal of medical systems, 34(5),pp.849858.Kuppusamy, P. and Divya, B., 2016. A SURVEY OF RETINA BASED DISEASEIDENTIFICATION USING BLOOD VESSEL SEGMENTATION. ICTACT Journal on Image &Video Processing, 7(2).Liu, I. and Sun, Y., 1993. Recursive tracking of vascular networks in angiograms based on thedetection-deletion scheme. IEEE Transactions on Medical Imaging, 12(2), pp.334-341.Lupascu, C.A., Tegolo, D. and Trucco, E., 2010. FABC: retinal vessel segmentation usingAdaBoost. IEEE Transactions on Information Technology in Biomedicine, 14(5), pp.1267-1274.Marín, D., Aquino, A., Gegúndez-Arias, M.E. and Bravo, J.M., 2011. A new supervised methodfor blood vessel segmentation in retinal images by using gray-level and moment invariantsbasedfeatures. IEEE Transactions on medical imaging, 30(1), pp.146-158.Mendonca, A.M. and Campilho, A., 2006. Segmentation of retinal blood vessels by combiningthe detection of centerlines and morphological reconstruction. IEEE transactions on medicalimaging, 25(9), pp.1200-1213.Osareh, A. and Shadgar, B., 2009. Automatic blood vessel segmentation in color images ofretina. Iranian Journal of Science and Technology, 33(B2), p.191.Rahebi, J. and Hardalaç, F., 2014. Retinal blood vessel segmentation with neural network byusing gray-level co-occurrence matrix-based features. Journal of medical systems, 38(8), p.85.Sample Script17Ricci, E. and Perfetti, R., 2007. Retinal blood vessel segmentation using line operators andsupport vector classification. IEEE transactions on medical imaging, 26(10), pp.1357-1365.Soares, J.V., Leandro, J.J., Cesar, R.M., Jelinek, H.F. and Cree, M.J., 2006. Retinal vesselsegmentation using the 2-D Gabor wavelet and supervised classification. IEEE Transactions onmedical Imaging, 25(9), pp.1214-1222Staal, J., Abràmoff, M.D., Niemeijer, M., Viergever, M.A. and Van Ginneken, B., 2004.Ridgebased vessel segmentation in color images of the retina. IEEE transactions on medicalimaging, 23(4), pp.501-509.Sum, K.W. and Cheung, P.Y., 2008. Vessel extraction under non-uniform illumination: a level setapproach. IEEE Transactions on Biomedical Engineering, 55(1), pp.358-360.Taylor, H.R. and Keeffe, J.E., 2001. World blindness: a 21st century perspective. British Journalof Ophthalmology, 85(3), pp.261-266.Taylor, H.R. and Keeffe, J.E., 2001. World blindness: a 21st century perspective. British Journalof Ophthalmology, 85(3), pp.261-266.Villalobos-Castaldi, F.M., Felipe-Riverón, E.M. and Sánchez-Fernández, L.P., 2010. A fast,efficient and automated method to extract vessels from fundus images. Journal of Visualization,13(3), pp.263-270.Wasan, B., Cerutti, A., Ford, S. and Marsh, R., 1995. Vascular network changes in the retinawith age and hypertension. Journal of hypertension, 13(12), pp.1724-1728.You, X., Peng, Q., Yuan, Y., Cheung, Y.M. and Lei, J., 2011. Segmentation of retinal bloodvessels using the radial projection and semi-supervised approach. Pattern Recognition, 44(10),pp.2314-2324.Zana, F. and Klein, J.C., 2001. Segmentation of vessel-like patterns using mathematicalmorphology and curvature evaluation. IEEE transactions on image processing, 10(7),pp.10101019.Zhang, B., Zhang, L., Zhang, L. and Karray, F., 2010. Retinal vessel extraction by matched filterwith first-order derivative of Gaussian. Computers in biology and medicine, 40(4), pp.438-445.Zhao, Y.Q., Wang, X.H., Wang, X.F. and Shih, F.Y., 2014. Retinal vessels segmentation basedon level set and region growing. Pattern Recognition, 47(7), pp.2437-2446.Zolfagharnasab, H. and Naghsh-Nilchi, A.R., 2014. Cauchy based matched filter for retinalvessels detection. Journal of medical signals and sensors, 4(1), p.1.18Appendix A – Ethics Form[Complete after approval]Department of Computer and Information SciencesSTUDENT PROJECT: ETHICS REGISTRATION AND APPROVAL FORMSection One: Registration [To be completed bystudent] Title of projectBlood vessel segmentation in retinal imagesResearcher’s nameProgramme of studyMSc computer scienceAcademic YearModule codeKF7028Supervisor’s nameAmmar BelatrecheSecond Marker’s nameShelagh KeoghStart date of project22/05/2017 Short description of the project, including research methods and selection of any participants: RedAmberGreen Sample Script19Automatic retinal blood vessel segmentation algorithms are important procedures inthe computer-aided diagnosis in the field of ophthalmology. They help to produceuseful information for the diagnosis and monitoring of eye diseases such as diabeticretinopathy, hypertension and glaucoma. Retinal images have been widely used fordiagnosing vascular and non-vascular pa-theology in the medical society. Retinalimages provide information on the changes in retinal vascular structure, which arecommon in diseases such as diabetes, stroke and cardiovascular disease. Thesediseases usually change reflectivity, tortuosity, and patterns of blood vessels. Forexample, hypertension changes the branching angle or tortuosity of vessels anddiabetic retinopathy can lead to neovascularization, i.e., development of new bloodvessels. If left untreated, these medical conditions can cause sight degradation oreven blindness. The early exposure to these changes is important for taking preventivemeasure and hence, the major vision loss can be preventedAutomatic segmentation of retinal blood vessels from retinal images would be apowerful tool for medical diagnostics. For this purpose, the segmentation method usedshould be as accurate and reliable as possible. The main aim of segmentation is todifferentiate an object of interest and the background from an image.I will evaluate existing blood vessel segmentation techniques and develop andeffective a more effective and efficient approach.The methods are:Image process techniquesFeatures extractionClassificationPost processing20 10. If yes [to 5, 6, 7, 8 or 9 above] have you identified steps to address the issues? Statement by researcherThis statement should explain how any issues identified in the answers to the above questions will be addressedand what steps will be taken to mitigate such risks or adverse impactThe project will not carry out any kind issues.I have read the University and the Faculty Ethics Policy and Procedures and confirm that the answers I havegiven above are correct. Where issues arise under items 5, 6, 7, 8 or 9 [above] I have described in writing howI intend to approach these issues in the research.Researcher’s signature Asmail MuftahDate 24/03/2017Sample Script21Section Two: Approval[The form is reviewed by the supervisor and second marker. Approval maybe given by either for green projects;amber projects must be approved by the second marker. Red projects must be referred to the Faculty ResearchEthics Committee.]Red: Vulnerable participants, sensitive data, risks to participants or researchers, NHS, etc.Amber: Human participants, environmental issues, commercially sensitive information, etc.Green: No participants involved, no sensitive data, etc.For full definitions see section on Risk Categories in the Engineering and Environment Ethics Procedures.Ethical approval[Please tick as appropriate] Green – Ethical approval is given without conditionsAmber – Ethical approval is given with the following conditionsInformation to be provided to all participantsParticipant consent to be obtained using the standard Research Participant ConsentForm or otherwise in accordance with Faculty proceduresData to be stored and destroyed securely in accordance with University guidelinesAdherence to Data Protection ActAnonymity to be provided to participantsCommercial confidentiality to be provided to organisations(s)Other (please state):Red – Project is referred to FREC for approval 22 Name of Approver……………………………………….Signature……………………………………….Date………………………………………. Outcome of FREC referral – Decision, minute and date of meeting, or signatures of two signatories, one ofwhom is a member of FREC.
