Computer Vision News - August 2021

3 Summary 17 AutomatedCartilage Segmentation the sampled cells for several weeks in a specialized lab, and implanting the grown cells in the damaged regions. For the physician to accurately assess the extent of the cartilage damage and select the most fitting conservational treatment, the cartilage must be imaged and segmented accurately and non-invasively. Articular cartilage is imaged using magnetic resonance imaging (MRI) and then segmented and evaluated for lesion shape, size (diameter), and boundaries. In addition, any healthy cartilage in non- weightbearing regions is mapped for possible cartilage harvesting and transfer. Cartilage segmentation using MRI The novel cartilage segmentation tool developed by RSIP Vision uses artificial intelligence technology to provide accurate assessment of articular cartilage health in MRI scans. This new module provides fully automated segmentation and precise assessment of the integrity of damaged cartilage in a variety of joints, such as hips, knees, and ankles. Using deep learning algorithms, RSIP Vision’s module accurately measures the exact location, diameter, shape, and boundaries of both chondral lesions as well as healthy, non-weight bearing cartilage. “It is very valuable to be able to accurately map chondral lesions preoperatively,” says Dr Shai Factor , an orthopedic surgeon at Tel Aviv Sourasky Medical Center in Israel. “Analyzing the parameters of the lesion and its boundaries allows the surgeon, along with the patient, to choose the ideal cartilage repair technique” . Evaluating these crucial geometric features of both healthy and damaged cartilage supports the physician in selecting the most suitable cartilage-sparing procedure , resulting in improved patient outcome and shortened recuperation downtime for athletes . RSIP Vision’s segmentation module offers not only an improved way to evaluate cartilage damage and select the best treatment, but also a way to follow-up postoperatively and assess the treatment efficacy. This module follows RSIP Vision’s novel module for segmenting bones in computerized tomography (CT) scans with metal implants present . This existing tool was designed and trained to deal with the presence of metals in CT scans, offering accurate and robust segmentation of both bones and implants. Together, these two novel modules offer better diagnosis and treatment for patients undergoing orthopedic joint procedures, including the many patients that undergo additional or follow-up procedures throughout their lifetime.