Please use this identifier to cite or link to this item: http://hdl.handle.net/10603/458767
Full metadata record
DC FieldValueLanguage
dc.coverage.spatialExponentially weighted and heaped multilayer perceptron for skin cancer diagnosis using multi directional systems
dc.date.accessioned2023-02-16T08:45:47Z-
dc.date.available2023-02-16T08:45:47Z-
dc.identifier.urihttp://hdl.handle.net/10603/458767-
dc.description.abstractSkin cancer is the most dangerous and deadliest disease and its incidence and mortality rate increases worldwide. The analysis of dermoscopic images is the most appropriate non-invasive method for skin cancer diagnosis as a large amount of diagnostic information can be gleaned from dermoscopic images. Though many existing algorithms available for feature extraction and classification for a particular medical image classification system, the main aim of study is to improve the accuracy of the skin cancer classification system. The selection of features and classifiers can greatly improve the accuracy of diagnosis and reduces the computation time. This research work aims to build an effective dermoscopic image classification system to diagnose skin cancer at the earliest. To achieve this goal, a Hybrid Artificial Intelligence Model (HAIM) is developed for dermoscopic image classification that combines different multi-directional representation systems with an efficient Exponentially Weighted and Heaped MultiLayer Perceptron (EWHMLP) for the successful classification.At first, the use of texture analysisis explored by the use of frequency domain analyzes to find out if the dermoscopic images can be classified from texture. Among the transformations techniques, Wavelets is a multiresolution analysis and Curvelet, Contourlet and Shearlet are the multidirectional analyzes. The proposed system has two stages; information or feature extraction from the dermoscopic images and then classification. Features which are discriminate in nature are extracted using different transformation in the former process and they are classified in the later stageusing the classifier. Two different texture features; energy and entropy are extracted from the sub-bands obtained by the decomposition process of each technique. In order to identify the best feature among the energy and entropy texture features, these features are extracted for different decomposition levels starting from 1 to 4. The extracted energy and entropy features are anal
dc.format.extentxvii,155
dc.languageEnglish
dc.relationp.147-154
dc.rightsuniversity
dc.titleExponentially weighted and heaped multilayer perceptron for skin cancer diagnosis using multi directional systems
dc.title.alternative
dc.creator.researcherVidyalakahmi, V
dc.subject.keywordEngineering and Technology
dc.subject.keywordComputer Science
dc.subject.keywordTelecommunications
dc.subject.keywordMulti directional systems
dc.subject.keywordMultilayer perceptron
dc.subject.keywordCurvelet
dc.description.note
dc.contributor.guideLeena Jasmine, J S
dc.publisher.placeChennai
dc.publisher.universityAnna University
dc.publisher.institutionFaculty of Information and Communication Engineering
dc.date.registered
dc.date.completed2022
dc.date.awarded2022
dc.format.dimensions21cm
dc.format.accompanyingmaterialNone
dc.source.universityUniversity
dc.type.degreePh.D.
Appears in Departments:Faculty of Information and Communication Engineering

Files in This Item:
File Description SizeFormat 
01_title.pdfAttached File19.58 kBAdobe PDFView/Open
02_prelim pages.pdf1.24 MBAdobe PDFView/Open
03_contents.pdf123.71 kBAdobe PDFView/Open
04_abstract.pdf123.43 kBAdobe PDFView/Open
05_chapter 1.pdf393.74 kBAdobe PDFView/Open
06_chapter 2.pdf304.33 kBAdobe PDFView/Open
07_chapter 3.pdf960.25 kBAdobe PDFView/Open
08_chapter 4.pdf768.95 kBAdobe PDFView/Open
09_chapter 5.pdf3.9 MBAdobe PDFView/Open
10_annexures.pdf110.95 kBAdobe PDFView/Open
80_recommendation.pdf71.17 kBAdobe PDFView/Open


Items in Shodhganga are licensed under Creative Commons Licence Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0).

Altmetric Badge: