Year: 2024
Type: conference-paper
Abstract: Manual and automatic segmentation techniques can be applied to DICOM medical images from magnetic resonance imaging (MRI) to extract certain structures, such as soft tissues, but the precise extraction of bone structures may be limited. This study studies these types of knee bone tissue segmentation on MRI, to avoid the need to resort to computed tomography (CT) for obtaining the desired bone structures. Manual segmentation was done using ITK-Snap and automatic segmentation algorithms were applied in Python and the ITK library. As a result of this study, it was found that although manual segmentation allowed for precise and consistent identification of the femur, tibia, fibula, and patella, the automatic segmentation needed to achieve the same level of accuracy.
Authors: Vasco Silva, Adélio Vilaça, Rita Veloso, Luís Coelho, Renato Magalhães
Institution: Universidade Técnica do Porto
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4407417903