Comparative Evaluation Framework for AI-Based Video Motion Capture Tools

Abstract

With the rapid adoption of AI-based video motion capture tools in digital animation, understanding their comparative performance has become essential. This study presents a comparative performance analysis and practical evaluation framework for AI-based video motion capture tools. Rokoko Vision, DeepMotion, and Meshcapade were assessed for their perceived motion-tracking quality, gesture recognition, motion stability, and usability under standardized conditions using a single-camera setup and five representative motions, namely rotation, walking, squatting, jumping, and hand gestures. DeepMotion exhibited stronger upper-body and gesture fidelity, Meshcapade produced more anatomically consistent full-body motion, and Rokoko Vision offered stable baseline tracking with high accessibility. The proposed framework provides workflow-specific insights, particularly for previsualization, small-studio production, and animation education, thereby helping creators to select tools that align with their production goals.