Please use this identifier to cite or link to this item: http://hdl.handle.net/10603/62161
Title: Machine Learning with Optimized Features for Enriched Performance of Content Based Image Retrieval
Researcher: Priya R
Guide(s): Vasantha Kalyani David
Keywords: Content Based Image Retrieval
Support Vector Method
Classification B+ tree
University: Avinashilingam Deemed University For Women
Completed Date: 08.10.2015
Abstract: Demand for sophisticated image retrieval is increasing in a tremendous fashion in both offline and online community Content Based Image Retrievalsystems (CBIR) are used extensively to answer these demands CBIR also known as query by image content and content based visual information retrieval is the newlineapplication of computer vision techniques to the image retrieval problem, that is the problem of searching for digital images in large databases The main aim of this research work is to design and develop techniques that improve the process of image retrieval in terms of accuracy and speed To achieve this research goal the newlineproposed CBIR systems is designed in four major steps They are preprocessing feature extraction and reduction model construction and query Each of these tasks is interrelated and the output from one phase is used as input to the subsequent phases Preprocessing step is concerned with addressing three subtasks They are newlinestandardizing the images normalization and color space conversion The standardization of images to a fixed size of 128 cross 128 pixels is performed using a wavelet based bicubic and edge sensitive interpolation technique A hybrid color space model that combines enhanced R Red G Green B Blue color components newlinefrom RGB color space H Hue color component from HSV color space and chroma Cb and Cr color components from YCbCr is proposed The RGB color components are enhanced to handle the issues like high correlation and inability to reflect actual human perceptual color distance in the conventional RGB color space The second step of the proposed CBIR system extracts three categories of features that are to be used during image retrieval They are histogram based color features 5 shape features 15 and texture 7 The color features extracted are from an enhanced histogram that combines spatial information and avoids the newlineproblem of being sensitive to colors located in the central image Initially a color quantization using enhanced KMeans algorithm is used to reduce the number of discriminate colo
Pagination: 
URI: http://hdl.handle.net/10603/62161
Appears in Departments:Department of Computer Science

Files in This Item:
File Description SizeFormat 
rpriya_chapter 1.pdfAttached File373.81 kBAdobe PDFView/Open
rpriya_chapter 2.pdf276.04 kBAdobe PDFView/Open
rpriya_chapter 3.pdf175.81 kBAdobe PDFView/Open
rpriya_chapter 4.pdf442.32 kBAdobe PDFView/Open
rpriya_chapter 5.pdf352.24 kBAdobe PDFView/Open
rpriya_chapter 6.pdf429.67 kBAdobe PDFView/Open
rpriya_chapter 7.pdf2.94 MBAdobe PDFView/Open
rpriya_chapter 8.pdf93.49 kBAdobe PDFView/Open
rpriya_chapter 9.pdf226.75 kBAdobe PDFView/Open
rpriya_intro.pdf192.75 kBAdobe PDFView/Open
Show full item record


Items in Shodhganga are protected by copyright, with all rights reserved, unless otherwise indicated.

Altmetric Badge: