Master thesis presentation by Grzegorz Sowulewski, MPBME
Welcome to a master thesis presentation: Limb movement tracking and analysis for neurological dysfunction of neonates by using multi-camera videos
Presented by Grzegorz Sowulewski, MPBME
Supervisor and examiner: Irene Gu
When: Friday, June 12th at 13.15
Where: Lunnerummet (room 3311), Hörsalsvägen 11, 3rd floor
Abstract Neurological diseases of newborn babies are a common problem. When diagnosed early, this problem can be mitigated through medical treatment. This project mainly focus on video analysis of limb movement from neonates. Videos were measured from Östra hospital using 3 web-cameras. In the thesis work, detection of markers is performed by SIFT with post-processing for enhancement. Epipolar geometry is employed to 2D markers in different views to 3D space, where camera calibration is done by using an existing software. Particle filters are used for tracking trajectories of detected markers. Finally, some preliminary analysis is performed for quantifying the movement through analyzing the limb movement trajectories. Performance of tracking is then evaluated by projecting the points from 3D tracked trajectories in one of the 2D image views. Manually selected points are then used as the ground truth. Visual observation and error analysis are also shown.