GitHub - thetechgirl14/Brain-CT-Hemorrhage-Classification-Segmentation: In this project, we constructed a classification model and a segmentation model for brain CT images based on CNN (Convolutional Neural Network) and U-Net architecture for CNN. Six classes are involved based on different categories of the hemorrhage in patient’s brain, whereas the segmentation aims to provide the exact location of the hemorrhage for a given brain CT image. We cropped the 3D data of brain CT scan such that it compactly fits the ventricular region. We used 3D U-net as semantic segmentation, with weighted binary crossentropy loss. Finally, we reach 68% accuracy to our CNN classification model and 92% accuracy to our U-net segmentation model for testing sets.

thetechgirl14 / Brain-CT-Hemorrhage-Classification-Segmentation Public

forked from anupam312nwd/MATH7243_CT-scan_Classification_Segmentation

In this project, we constructed a classification model and a segmentation model for brain CT images based on CNN (Convolutional Neural Network) and U-Net architecture for CNN. Six classes are involved based on different categories of the hemorrhage in patient’s brain, whereas the segmentation aims to provide the exact location of the hemorrhage …

0 stars 1 fork Branches Tags Activity

Star

Notifications

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 2 Commits
Hemorrhage_Classification		Hemorrhage_Classification
Hemorrhage_Segmentation		Hemorrhage_Segmentation
MRI-3D-Segmentation		MRI-3D-Segmentation
Machine_Learning_Project_Report.pdf		Machine_Learning_Project_Report.pdf