Journal Article

·2020

3D Femoral Head Detection in MRI Data Sequences with the Integro-differential Operator

Abbas Memiş YTU , Songul Varli YTU , Fuat Bilgili

2020 Medical Technologies Congress (TIPTEKNO)

Abstract

This paper introduces a study of automatic femoral head detection in magnetic resonance imaging (MRI) data sequences. For the 3D detection of the multiform femoral heads having both spheric and aspheric shape structures, the threedimensional form of the Integro-differential Operator (IDO) was performed. Following a set of image pre-processing operations including image intensity normalization, histogram equalization, morphological correction, hip joint separation and image binarization performed on bilateral hip MRI data sequences, the hip joints images are obtained in binary form in 3D. Then, the 3D form of IDO is performed in a predefined image volume to detect the femoral heads. Within the experimental studies performed on 8 bilateral hip MRI data sequences belonging to 6 LeggCalve-Perthes disease (LCPD) patients, promising success rates were observed. In detection of a total of 16 femoral heads, 8 of which are spheric and 8 of which are aspheric, 0.7021 (± 0.3160) and 0.6757 (± 0.2989) DSC values measured for the spheric and aspheric femoral heads, respectively.

Keywords

Femoral head Magnetic resonance imaging Artificial intelligence Histogram equalization Computer vision Image processing Nuclear medicine Computer science Medicine Radiology Image (mathematics) Anatomy

Subject Areas

Hip disorders and treatments ·Surgery ·Health Sciences
Cardiac Valve Diseases and Treatments ·Cardiology and Cardiovascular Medicine ·Health Sciences

Citations by Year