Rokoko

🚨 Breaking: Rokoko is solving one of robotics' biggest data collection problems. The challenge is not the robot. It is not the AI model. It is occlusion. Most teleoperation systems depend on cameras to track human motion. The moment an operator reaches inside a box, behind an object, or into a machine, tracking quality drops because the cameras lose sight of the hands. For imitation learning, that means incomplete demonstrations and lower-quality training data. Rokoko takes a different approach. Their Smartgloves II combine high-rate IMUs with an electromagnetic tracking system, enabling accurate finger and hand tracking without relying on camera line of sight. The result: • Continuous tracking even when hands are completely hidden • Precise finger motion capture • Reduced drift over long sessions • Higher-quality datasets for robot learning As physical AI and imitation learning continue to scale, breakthroughs will not come from larger models alone. They will come from better data. And that starts with better motion capture. ⚡ Learn more: https://www.rokoko.com/ ⚡ Join 4,700+ robotics engineers: https://lnkd.in/dYxB9iCh ⚡ Subscribe for weekly robotics insights: https://lnkd.in/eVEnffCJ #Robotics #Teleoperation #ImitationLearning #PhysicalAI #EmbodiedAI #MotionCapture

A little insight into how I went about animating The Maxx. Everyone loves a BTS and these days you really have to show your workings to prove you're not cheating with that new supercomputer "thing" that everyone's talking about. I used keyframes for the Maxx and motion capture for the drunk dude, so best of both worlds. With Cultlovesyou Rokoko DAHOUSE AUDIO Final Frontier #animation #mocap #characteranimation #animatedshort

Animators and robotics engineers are discovering they have the same unsolved problem. How do you accurately capture a human hand interacting with a real object, while keeping every finger, joint, and prop in perfect alignment? Optical systems struggle with occlusion: the moment a hand wraps around a prop or moves out of camera view, you lose data. IMU systems without global position have a different problem: the hand and prop drift apart in virtual space, and aligning them in post is time-consuming and imperfect. In this video, our creative director Sam demonstrates what it looks like when both problems are solved at once. Smartsuit Pro II, Smartgloves II, Coil Pro. Unsheathing, striking, regripping. Every joint, finger, and the prop in perfect alignment throughout. No corrections needed. The Coil Pro's global position reference transfers directly to the Smartsuit Pro II, so the entire body (suit, gloves, fingers) shares one absolute spatial reference from the moment of capture. The same setup that gives animators flawless prop interaction is now giving robotics teams clean manipulation data. Wearable, real-time, no line-of-sight required, accurate down to every finger. 🗡️🤖

One of the things I enjoy most about creating cinematics is seeing the transformation from acting to the final character. Scene created in Unreal Engine + Motion Builder, combining motion capture for the first half with hand-keyed animation for the second. I don't believe an animator should rely exclusively on either technique, the real beauty lies in finding the right balance between the two. Motion Capture performed using a Rokoko Smartsuit Pro II from Frame Games Music licensed from Instagram: @harpazz_official

Full-body motion capture meets detailed finger tracking.   With the Rokoko Gloves Plugin for Motive, teams can bring Rokoko Smartgloves data directly into their OptiTrack workflow - combining precise optical tracking with expressive hand and finger performance.   See the setup guide on OptiTrack or Rokoko's website. OptiTrack: https://lnkd.in/gqQXWUrE Rokoko: https://lnkd.in/gxGm_qq5 #OptiTrack #Rokoko #MotionCapture #Mocap #FingerTracking