Comprehensive Investigation on Content-based Medical Image Retrieval using Radon Barcodes

Content Based Medical Image Retrieval (CBMIR) systems are vital to the underlying operation of medical databases because they allow quick search and retrieval of medical images. Radon Barcode (RBC)s are binary complementary feature vectors which increase the speed of CBMIR systems through smaller feature vector size and low retrieval error. We explore further improving the efficiency and accuracy of RBCs by optimizing the way they are extracted from medical images. Through the addition of image pre-processing, novel barcoding techniques, and improved distance evaluation methods, we improved RBC utility in CBMIR applications. Image pre-processing techniques such as histogram equalization and adaptive thresholding reduced the retrieval error of generated RBCs. We also introduce several novel barcode generation techniques such as Binary Coded Decimal Radon Barcodes (BCDRBC), Difference of Radon Projections Barcodes (DRPBC), and Difference of Radon Projections Soft Hash Barcode (DRPSHBC) which decreased both retrieval error and barcode size.

Name thesis:degree_name

Master of Applied Science (MASc)

Discipline thesis:degree_discipline

Electrical and Computer Engineering

Grantor

University of Ontario Institute of Technology

Year dc:date.issued

2019

Author dc:creator

• Rizvi, Syed Raazi

Advisor dc:contributor.advisor

• Rahnamayan, Shahryar

• Content-based medical image retrieval
• Radon barcodes
• Image processing

Language dc:language.iso

en

Handle dc:identifier.uri

https://hdl.handle.net/10155/1139

OAI identifier oai:identifier

oai:ontariotechu.scholaris.ca:10155/1139