AUTOMATIC CLASSIFICATION OF BRAIN TUMOR GRADE IN MRI IMAGE USING CONVOLUTIONAL NEURAL NETWORK (CNN) METHOD
Malignant brain tumor or brain cancer is a third deadly cancer. The accurate diagnosed of brain tumor grade is very important to make a treatment recommendation. An automatic classification tools as an aided tools is expected to reduce a human error diagnosed. The current study is evaluate automatic...
Saved in:
| Main Author: | |
|---|---|
| Format: | Theses |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/60068 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Malignant brain tumor or brain cancer is a third deadly cancer. The accurate diagnosed of brain tumor grade is very important to make a treatment recommendation. An automatic classification tools as an aided tools is expected to reduce a human error diagnosed. The current study is evaluate automatic classification method for four different grade of brain tumor (grade II, grade III, grade IV and non tumor) using Convolutional Neural Network with three kind of architecture, namely own-making architecture, Resnet-152 and VGG-16. The datasets used in this study are REMBRANDT dataset for brain tumor grade II, III and IV and IXI dataset for healthy or non-tumor brain. The number of image for each dataset are increased using
augmentation technique. The result show all of architectures are good in predict brain tumor grade with accuracy of 84%, 95% and 84% for own-making architecture, Resnet-152 and VGG-16 respectively. Based on these result, resnet-152 become the best CNN model for predict brain tumor grade. |
|---|