Computerized 2D detection of the multiform femoral heads in magnetic resonance imaging (MRI) sections with the integro-differential operator


Memis A., Varlı S. , Bilgili F.

BIOMEDICAL SIGNAL PROCESSING AND CONTROL, cilt.54, 2019 (SCI İndekslerine Giren Dergi) identifier identifier

  • Cilt numarası: 54
  • Basım Tarihi: 2019
  • Doi Numarası: 10.1016/j.bspc.2019.101578
  • Dergi Adı: BIOMEDICAL SIGNAL PROCESSING AND CONTROL

Özet

In this paper, an approach for the automatic 2D detection of the multiform femoral heads in magnetic resonance imaging (MRI) sections with the integro-differential operator (IDO) is presented. In the proposed study, MRI sections were firstly preprocessed to enhance the MR image quality. In the first phase of the preprocessing, MR image intensity normalization, histogram equalization and regional image filling were performed on the input MRI sections. In the second phase, left and right hip joints in bilateral hip images in coronal plane were separated equally in vertical axis. Then, colors of the MR images were quantized and separated hip images were represented as binary images. Finally, femoral heads were detected by performing the IDO. Experimental results observed on the spheric and aspheric femoral heads in 24 MRI sections of the 13 patients with Legg-Calve-Perthes disease (LCPD) show that proposed approach has a successful performance in detection of the multiform femoral heads. Proposed approach can be also used in determination of the initial point in automatic or semi-automatic segmentation, determination of fiducial points in measurement of the clinical parameters except for the detection. Femoral head detection with IDO was performed on coronal MRI sections in the proposed study, but it could also be performed on MRI sections acquired in other planes such as axial and sagittal. Moreover, this method could be used to detect the femoral heads or the other spheric anatomical structures such as humeral heads in other medical imaging modalities such as computed tomography (CT). (C) 2019 Elsevier Ltd. All rights reserved.