The paper presents a Bayesian approach for tracking 3D articulated human figures in monocular video sequences. The method defines a generative model of image appearance, a robust likelihood function based on image graylevel differences, and a prior probability distribution over pose and joint angles. The posterior probability distribution is represented using a discrete set of samples and propagated over time using particle filtering. This approach extends previous work by incorporating a perspective camera model, modeling limb self-occlusion, and recovering 3D motion from monocular sequences. The method handles ambiguities due to image matching, model singularities, and perspective projection, making it suitable for tracking under changing viewpoints, in grayscale images, and with complex backgrounds. The paper also reviews related work on image cues, body representations, temporal models, and estimation techniques, highlighting the advantages of the proposed method over existing approaches.The paper presents a Bayesian approach for tracking 3D articulated human figures in monocular video sequences. The method defines a generative model of image appearance, a robust likelihood function based on image graylevel differences, and a prior probability distribution over pose and joint angles. The posterior probability distribution is represented using a discrete set of samples and propagated over time using particle filtering. This approach extends previous work by incorporating a perspective camera model, modeling limb self-occlusion, and recovering 3D motion from monocular sequences. The method handles ambiguities due to image matching, model singularities, and perspective projection, making it suitable for tracking under changing viewpoints, in grayscale images, and with complex backgrounds. The paper also reviews related work on image cues, body representations, temporal models, and estimation techniques, highlighting the advantages of the proposed method over existing approaches.