M.Sc. Tezi Görüntüleme | |||||||||||||||||||||
|
|
||||||||||||||||||||
| Summary: In medicine, diagnosis of any disease is very important process as well as treatment. A successful treatment can only be done with accurate diagnosis. With respect to this, radiological imaging methods have important role in health care. Recently, Magnetic Resonance (MR) as a modern imaging method is often used especially for diagnosing the problems regarding knee joint. Meniscuses are cartilage cushions, which assume critical functions like supporting the knee joint, increasing the mobility by decreasing the burden that rides on the bones and cartilages. Tears that might be take place on this soft cartilage tissue are also very significant injury from the health of the knee point of view.In this study, first of all, the knee joint (meniscal area) which includes anterior and posterior horns of the medial or lateral meniscus are detected by histogram and statistical bone segmentation based methods on the sagittal plane, proton density MR images of the knee. Following several pre-processing operations on the localized knee joint, the image is converted to binary and labeled. Finally, template-matching technique is employed by means of adapting a parametric triangular prototype. Then, the meniscus horns are detected and analyzed for tear-diagnosis purpose. The diagnosis-analyze system has been tested on 100 slices of proton density, sagittal MR images of the knee that had been taken from 30 different patients, provided by Radiology Department of Faculty of Medicine at Karadeniz Technical University.
Keywords: Meniscus Tear, MR, Magnetic Resonance, Statistical Image Processing, Histogram, Template Matching, Pattern Recognition. | |||||||||||||||||||||