Human gait analysis is an important research field, as the analyzed markers are health, morbidity, survival, and mortality indicators. Important gait parameters are, for example, gait speed, clearance, or step lengths. However, the existing tools for capturing gait parameters can only be operated by trained professionals in controlled laboratories with expensive equipment. By creating a computer vision algorithm that can retrieve human movements of sufficient quality from 2D monocular images, patients' health can be tracked solely through handheld devices. This Master's Thesis aims to develop a human motion analysis system capable of detecting 3D skeletons from 2D images fitted to the task of human gait analysis. The system will be evaluated versus the marker trajectories of a VICON 3D motion capture system.