Use of novel interactive input devices for segmentation of articular cartilage from magnetic resonance images

Abstract

To study the effect of new interactive computer input devices on cartilage segmentation in terms of time, consistency between input devices, and precision in quantitative magnetic resonance imaging (qMRI). We compared two new input devices, an interactive digitizing tablet and an interactive touch-sensitive screen, to a traditional mouse. Medial tibial and patellar cartilage of six healthy and six osteoarthritic knees were segmented using each input device. Cartilage volume, surface area and mean thickness were assessed using a validated algorithm and used to determine consistency and precision. Segmentation time was also measured. Segmenting with an interactive touch-sensitive screen reduced segmentation time by 15% when compared to the traditional mouse but we found no significant difference in segmentation time between the interactive digitizing tablet and the traditional mouse. We found no difference in consistency or precision of cartilage volume, mean thickness or surface area between the three input devices tested. We conclude that measurements of cartilage made using articular cartilage segmentation from MR images are independent of the input device chosen for user interaction.