Please use this identifier to cite or link to this item:
http://hdl.handle.net/10603/361181
Title: | Design And Implementation Of A Hybrid Approach For The Optimization Of 3d Mri And Ct Brain Image Segmentation With Fusion Process |
Researcher: | Sumithra,M |
Guide(s): | Malathi,S |
Keywords: | Computer Science Computer Science Artificial Intelligence Engineering and Technology |
University: | Sathyabama Institute of Science and Technology |
Completed Date: | 2021 |
Abstract: | Recently, medical imaging field has established a widespread application in medical management for disease diagnosis, treatment preparation and disease monitoring which involves image acquisition of the organ affected by diverse methods. Application of several medical imaging is been carried out by the process of image segmentation which partitions an image into numerous divisions. Magnetic Resonance Imaging (MRI) brain images with precise segmentation assists the general practitioner efficiently in early diagnosis and also in brain tumour detection. Brain tumour, also known as intra cranial neoplasm is the abnormal growth of cells within the brain structure. It can be categorised into two types namely: Malignant and Benign. Benign tumour does not grow immediately and it hardly affects its adjacent healthy tissues. However, malignant tumour is very severe which tends to grow and worsen the condition in a span of time thereby being fatal. The MRI brain images are prone to inhomogeneous intensity of noise that invariably affects the pixel intensities. This in turn affects the precise spotting of the tumour cells leading to inaccurate segmentation results. It is understood from the literature survey that the existing segmentation algorithms are usually dependent on the uniformity of the image intensities and therefore is unsuccessful in providing accurate segmentation results in the MR brain images leading to false location of tumour cells. In this research, the focus is on a hybrid approach where algorithms are developed for different modalities. In first phase, a new approach called Distribution Matching Global and Local Fuzzy Clustering (DMGLC) is implemented for a single modality 3D brain MRI image for segmentation and classification using Dense AlexNet Neural Network incorporated with Region of Interest (ROI) as a hybrid model. This approach performs well in giving more newline newline newline newline newlineaccuracy but lags in handling two modalities. Nowadays for better diagnosis, two modalities are |
Pagination: | A5 |
URI: | http://hdl.handle.net/10603/361181 |
Appears in Departments: | COMPUTER SCIENCE DEPARTMENT |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
01. title.pdf | Attached File | 280.26 kB | Adobe PDF | View/Open |
02. certificate.pdf | 1.67 MB | Adobe PDF | View/Open | |
03. acknowledgement.pdf | 152.92 kB | Adobe PDF | View/Open | |
04. abstract.pdf | 200.14 kB | Adobe PDF | View/Open | |
05. table of contents.pdf | 748.69 kB | Adobe PDF | View/Open | |
06. chapter 1.pdf | 2.52 MB | Adobe PDF | View/Open | |
06. chapter 2.pdf | 3.59 MB | Adobe PDF | View/Open | |
06. chapter 3.pdf | 488.28 kB | Adobe PDF | View/Open | |
06. chapter 4.pdf | 1.59 MB | Adobe PDF | View/Open | |
06. chapter 5.pdf | 2.38 MB | Adobe PDF | View/Open | |
06. chapter 6.pdf | 2.08 MB | Adobe PDF | View/Open | |
07. conclusion.pdf | 27.07 kB | Adobe PDF | View/Open | |
08. references.pdf | 3.67 MB | Adobe PDF | View/Open | |
09. curriculam vitae.pdf | 5.25 kB | Adobe PDF | View/Open | |
10. evaluation reports.pdf | 2.75 MB | Adobe PDF | View/Open | |
80_recommendation.pdf | 280.26 kB | Adobe PDF | View/Open |
Items in Shodhganga are licensed under Creative Commons Licence Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0).
Altmetric Badge: