This paper introduces a novel visual motion estimation technique that accurately recovers high-degree-of-freedom articulated human body configurations in complex video sequences. The method integrates the product of exponential maps and twist motions into a differential motion estimation framework, enabling robust recovery of kinematic degrees-of-freedom even in noisy and self-occluded configurations. The authors demonstrate the effectiveness of this approach on several image sequences of people performing full-body movements, re-animating an artificial 3D human model, and recovering the famous movements from Eadweard Muybridge's motion studies. This is the first computer vision-based system capable of processing such challenging footage with high accuracy. The technique is based on solving simple linear systems, which simplifies the estimation process and enhances tracking robustness. The paper also discusses the integration of multiple camera views and adaptive support maps to improve the accuracy and reliability of the motion estimation.This paper introduces a novel visual motion estimation technique that accurately recovers high-degree-of-freedom articulated human body configurations in complex video sequences. The method integrates the product of exponential maps and twist motions into a differential motion estimation framework, enabling robust recovery of kinematic degrees-of-freedom even in noisy and self-occluded configurations. The authors demonstrate the effectiveness of this approach on several image sequences of people performing full-body movements, re-animating an artificial 3D human model, and recovering the famous movements from Eadweard Muybridge's motion studies. This is the first computer vision-based system capable of processing such challenging footage with high accuracy. The technique is based on solving simple linear systems, which simplifies the estimation process and enhances tracking robustness. The paper also discusses the integration of multiple camera views and adaptive support maps to improve the accuracy and reliability of the motion estimation.