Possibilities of texture based motion analysis

Analysis of motion patterns in the human locomotion apparatus is important in many clinical areas like orthopaedics, physiotherapy, neurology, and sports medicine. Today marker based human motion analysis (HMA) is completely dominant in the clinical context. Technically, these systems are stable and dependable, and about a dozen variants are commercially available. One drawback of such systems is the time consuming and error prone marker placement. The purpose of the present contribution is to show that it is possible, with existing simple technology and methods, to build systems that do not depend on anatomically placed markers and yet produce an accuracy in knee joint center estimation comparable to marker based systems. It is shown that texture based methods can give estimates of knee joint center of rotation that can be compared to knee joint center estimates from marker based systems. Due to different definitions of the knee joint center position some bias is seen between the two estimates.