Computer Vision News - April 2021

which result in the surrounding bright streaks and dark streaks between the metals, to Poisson noise and edge-related artifacts, due to under sampling, subject movement, cone beam, and windmill artifacts. These metal artifacts are particularly pronounced in iron and platinum, which have a high atomic number, and less pronounced in titanium and other low atomic number metals. The result is a challenging CT image for standard segmentation tools, which result in inaccurate segmentation that limit the ability of the orthopedic surgeon to properly plan the arthroplasty. RSIP Vision has launched a new 13 Metal Implant Segmentation The reason for these artifacts is that CT images consist of multiple X-ray images, taken using a rotating X-ray generator and detector. The images are acquired at multiple stages and frommultiple angles along the rotation axis. This rotation process causes the metal, which almost completely absorbs X-rays, to repeatedly create the artifacts, resulting in the star- shaped streak artifact. In general terms, metal artifacts are caused by nonstandard absorption values (Hausdorff) due to crosstalk between absorbing pixels and additional artifacts. These metal artifacts are caused by a variety of reasons, from beam hardening and scatter effects, Knee joint segmentation 1: A knee CT scan with image artifacts due to an existing metal implant (left) segmented to all bone and metal components (right)