الفهرس 
AbstractOnly 14 pages are availabe for public view
 |  | from | 125 |  |  |
| | | from | 125 | | |
Abstract
Medical imaging refers to the techniques and processes used to create images of the human body for various clinical purposes. CT is quite useful for doctors to analyze the pathological changes of the biological organs. This work presents a CAD system used to detect and classify brain hemorrhage grade.. In this thesis, due to the low contrast of CT brain DICOM images and noise in images, the image firstly preprocessed to improve visual appearance.Then we need to find ROI of brain hemorrhage by using an automatic method with tolerance range 0.7, 0.916, it was achieved accuracy up to 97.4%. Because of the different types noise that DICOM images suffers from such as Speckle and Gaussian noise comparison evaluation between different types of filters according to image quality metrics, such as PSNR, we found that the best filters for all types of noise is bilateral filter. In the next stage, brain segmentation algorithms have been done using active contour with thresholding operation to extract hemorrhage region from CT brain image. After that classification process is coming to classify hemorrhage grade using ANN with 7 input, 15 hidden neurons and 3 output present grade1, grade 2, grade 3 and the classification recognition rate reach to 95.6%.