Recently, significant advances have been made in many sub-areas regarding the problem of markerless human motion capture. However, current markerless systems tend to introduce major simplifications, especially in early stages of the process, that temper the robustness and the generality of any subsequent modules and, consequently, of the whole application. This paper concentrates on improving the aspects of multi-camera system design, multi-camera calibration and shape-from-silhouette reconstruction. A thoughtful system analysis is first proposed with the objective of achieving an optimal synchronized multi-camera system. This multi-camera system is then accurately calibrated using a flexible method which allows free camera positioning. A novel region-based silhouette extraction procedure is proposed to remove the requirement of static and highly contrasting backgrounds. The outcome of our work is the achievement of robust voxel data reconstruction and coloring in complex and unconstrained scenes.