Please use this identifier to cite or link to this item:
Title: Statistical Modelling, and Characterization of Brain Tumors
Authors: Iqbal, Sajid
Keywords: Computer Science
Computer & IT
Issue Date: 2020
Publisher: University of Engineering & Technology, Lahore.
Abstract: With the exponential developments in Artificial Intelligence, medical image pro cessing has also seen a dramatic expansion. Medical image processing is an in terdisciplinary field that encompasses computer science, mathematics, statistics, biology, physics and medicine. Computer aided diagnosis and intervention is the major application of medical image processing. With the development of new imaging modalities more and more information can be extracted from medical images i.e. MRI and availability of such multi-dimensional information poses new challenges and opens new vistas for researchers and practitioners. This thesis in tends to develop methods for the analysis, segmentation and classification of brain tumor images. A plethora of binary segmentation and classification is present in literature however comparatively little work is present for multi-classification where more than one class tumors are present in images. The problem is still an open area of research due to ambiguous properties of tumors and their overlap ping features. Artificial Intelligence has moved from simple logic programming to complex statistical methods i.e. machine learning. A specialized set of statisti cal methods known as deep learning has got lot of attraction from research and development community due to their high performance over traditional machine learning methods. A variety of neural networks are being explored and new models are being developed. In this thesis, we design high performance methods to solve the problem under focus using bench mark dataset.
Gov't Doc #: 22553
Appears in Collections:PhD Thesis of All Public / Private Sector Universities / DAIs.

Files in This Item:
File Description SizeFormat 
sajid iqbal cs 2020 uet lhr.pdfphd.Thesis3.04 MBAdobe PDFView/Open

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