Saygılı, AhmetAlbayrak, Songül2022-05-112022-05-112018978-1-5386-1501-02165-0608https://hdl.handle.net/20.500.11776/611026th IEEE Signal Processing and Communications Applications Conference (SIU) -- MAY 02-05, 2018 -- Izmir, TURKEYAutomatic segmentation and classification studies in medical images have been intensely studied in recent years. The results obtained will support the decisions of medical experts. In this study, features were obtained by applying histogram of oriented gradients (HOG) method to segmented knee MR images with fuzzy clustering approaches and these features were trained with different classifiers to perform automatic meniscus tear detection. For this automatic detection, 28 different MR images provided by the Osteoarthritis Initiative were used. In particular, the effects of HOG have been studied in detail. Support vector machines, extreme learning machines, and k-nearest neighbor classifiers have been used in the classification stage. The support vector machines became the most successful classifier with a success rate of 88.78%. It is aimed to increase the success of the system with different feature extraction and segmentation methods in the following studies.trinfo:eu-repo/semantics/closedAccessMeniscus tearsSegmentationClassificationMedical Image ProcessingHistogram of Oriented GradientsSegmentationConference ObjectN/AWOS:0005114485002282-s2.0-85050791523N/A