Brain tumor image segmentation using deep learning approach
The role of computer vision in the field of biomedical sciences is crucial. In neurosurgical, Magnetic Resonance Images (MRI) scans are used to detect cancerous cell grow thin brain called brain tumor. Application that can aid in providing automatic brain tumor segmentation is crucial as segmentatio...
| Main Author: | |
|---|---|
| Format: | Final Year Project / Dissertation / Thesis |
| Published: |
2022
|
| Subjects: | |
| Online Access: | http://eprints.utar.edu.my/4645/ http://eprints.utar.edu.my/4645/1/fyp_CS_2022_SD.pdf |
| _version_ | 1848886205002809344 |
|---|---|
| author | Darshan, Suresh |
| author_facet | Darshan, Suresh |
| author_sort | Darshan, Suresh |
| building | UTAR Institutional Repository |
| collection | Online Access |
| description | The role of computer vision in the field of biomedical sciences is crucial. In neurosurgical, Magnetic Resonance Images (MRI) scans are used to detect cancerous cell grow thin brain called brain tumor. Application that can aid in providing automatic brain tumor segmentation is crucial as segmentation of the exact size and spatial location of these tumors are a time-consuming task. In addition, the existing software for brain tumor segmentation are implemented using conventional image segmentation algorithms which are labor intensive and causes over or under segmented tumor regions. Deep learning algorithm is able to provide good tumor segmentation results compared to other conventional segmentation algorithms as it learns from the labeled brain MRIs to predict the location of tumor region and consequently segment the tumor. Therefore, in this work, deep learning approach is used to build unassigned ensemble model which automatically segments tumor regions. The proposed model is also deployed as a web application using Flask for public usage. The proposed model is developed with Keras by using “Brain MRI segmentation” dataset. 1373images are used from the dataset where they are split as 1098 images for training, 137imagesfor validation and 138 images for testing and the proposed model managed to obtain 75.78% average tumor segmentation accuracy using Jaccard similarity index on the testing set. The proposed ensemble model was also tested on another dataset with 621 testing images where it achieved 66.75% testing accuracy by using the Jaccard similarity index. |
| first_indexed | 2025-11-15T19:34:47Z |
| format | Final Year Project / Dissertation / Thesis |
| id | utar-4645 |
| institution | Universiti Tunku Abdul Rahman |
| institution_category | Local University |
| last_indexed | 2025-11-15T19:34:47Z |
| publishDate | 2022 |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | utar-46452023-01-16T13:55:13Z Brain tumor image segmentation using deep learning approach Darshan, Suresh Q Science (General) T Technology (General) The role of computer vision in the field of biomedical sciences is crucial. In neurosurgical, Magnetic Resonance Images (MRI) scans are used to detect cancerous cell grow thin brain called brain tumor. Application that can aid in providing automatic brain tumor segmentation is crucial as segmentation of the exact size and spatial location of these tumors are a time-consuming task. In addition, the existing software for brain tumor segmentation are implemented using conventional image segmentation algorithms which are labor intensive and causes over or under segmented tumor regions. Deep learning algorithm is able to provide good tumor segmentation results compared to other conventional segmentation algorithms as it learns from the labeled brain MRIs to predict the location of tumor region and consequently segment the tumor. Therefore, in this work, deep learning approach is used to build unassigned ensemble model which automatically segments tumor regions. The proposed model is also deployed as a web application using Flask for public usage. The proposed model is developed with Keras by using “Brain MRI segmentation” dataset. 1373images are used from the dataset where they are split as 1098 images for training, 137imagesfor validation and 138 images for testing and the proposed model managed to obtain 75.78% average tumor segmentation accuracy using Jaccard similarity index on the testing set. The proposed ensemble model was also tested on another dataset with 621 testing images where it achieved 66.75% testing accuracy by using the Jaccard similarity index. 2022-04-22 Final Year Project / Dissertation / Thesis NonPeerReviewed application/pdf http://eprints.utar.edu.my/4645/1/fyp_CS_2022_SD.pdf Darshan, Suresh (2022) Brain tumor image segmentation using deep learning approach. Final Year Project, UTAR. http://eprints.utar.edu.my/4645/ |
| spellingShingle | Q Science (General) T Technology (General) Darshan, Suresh Brain tumor image segmentation using deep learning approach |
| title | Brain tumor image segmentation using deep learning approach
|
| title_full | Brain tumor image segmentation using deep learning approach
|
| title_fullStr | Brain tumor image segmentation using deep learning approach
|
| title_full_unstemmed | Brain tumor image segmentation using deep learning approach
|
| title_short | Brain tumor image segmentation using deep learning approach
|
| title_sort | brain tumor image segmentation using deep learning approach |
| topic | Q Science (General) T Technology (General) |
| url | http://eprints.utar.edu.my/4645/ http://eprints.utar.edu.my/4645/1/fyp_CS_2022_SD.pdf |