Virdyn YanDong 7 Camera Markerless Motion Capture

Unbox performance without limits with the Virdyn YanDong — a complete, professional markerless motion capture system that liberates performers from suits and sensors. Adopting a completely wearable-free design, the YanDong system requires no motion capture suits, helmets, or physical markers throughout the entire capturing process. It utilizes multi-camera synchronous video acquisition to perform all round tracking of the human body, delicate fingers, and facial expressions. Powered by advanced artificial intelligence algorithms, the system conducts real time AI human detection and 3D human reconstruction, followed by precise AI skeleton motion solving combined with forward and inverse kinematics calculation logic and a temporal transformer network for intelligent motion correction and frame to frame smoothing, delivering stable, jitter-free, and highly natural motion output.

Why Choose the Virdyn YanDong System?

• Truly Markerless – No suits, no markers. Just pure, unhindered performance capture for one or two actors simultaneously in a 4m × 4m area.
• Comprehensive Capture – Achieve holistic animation data with precise tracking of body, intricate hand movements, and facial expressions all at once.
• Seamless Pipeline Integration – Real-time data streaming via UDP and dedicated plugins for all major platforms including Unity, Unreal Engine, Maya, Blender, MotionBuilder, and iClone, with export in FBX, BVH, and CSV formats.
• Professional-Grade Hardware – 7x 3.2MP color global shutter cameras with ultra wide angle, low distortion lenses ensure high-quality, synchronized data capture at 60 fps.
• All-in-One Package – Everything needed in one box cameras, calibration tools, software, and full support for a quick and easy setup.

Precision Hardware — 7-Camera Array (VIRDYN YDC-V2):

Seven professional grade industrial cameras form the foundation of the YanDong system’s accuracy, each featuring a 3.2MP color CMOS global shutter sensor, C mount ultra wide angle lenses with exceptionally low distortion (<0.5%), an 82°×66° field of view, and Gigabit Ethernet (GigE Vision Protocol) with Power over Ethernet (PoE) support. Rated for operation from -30°C to 60°C, each camera is clearly labeled for quick identification and setup.

Powerful Software — YanDongFang:

The proprietary YanDongFang software leverages advanced AI image recognition to transform video feeds into precise 3D skeletal data:

• Comprehensive Skeletal Tracking – Real time capture of 21 major body joints (including displacement and quaternion data), 30 individual finger joints for natural hand animations, and simultaneous facial expression capture for full emotional performance.
• Effortless Camera & System Management – One click synchronization and control for all 7 cameras, with adjustable resolution, frame rate, exposure, and gain, plus simple intrinsic/extrinsic calibration using a checkerboard or calibration rod.
• Flexible Capture Modes – Choose between Single Actor or Dual Actor mode, toggle finger and facial capture on/off, and capture within a volume of up to 4m × 4m.
• Real-Time Data & Visualization – Ultra low latency of less than 80ms from performance to digital output, with real time preview of all camera feeds and live 3D model visualization (skeleton or robot model) that can be rotated, panned, and zoomed.
• Robust Data Handling & Export – Real time synchronized multi-camera video recording, real time data streaming via UDP broadcast, and export in FBX, BVH, CSV, and TXT formats.

4 Core Industry Applications:

• Motion Analysis with VDAnalyze – Combine with Virdyn’s professional motion analysis software for dance training, sports performance optimization, and rehabilitation assessment with real-time feedback.
• Digital Human Driving for Exhibition Halls – Control virtual digital humans naturally and in real time, completely markerless and suit free ideal for digital showrooms, brand presentations, and museums.
• Offline Exhibition & VR Game Interaction – Create engaging motion interactive experiences for trade shows and entertainment venues, with visitors participating in VR motion games and real time body interaction.
• VLA Motion Data Collection – Capture high quality human motion data efficiently and precisely, supporting realistic animation generation, live virtual production, AI training datasets, and virtual event broadcasting.