Hand Tracking V5: Gemini Developer Preview!
Posted; January 27, 2021
A lot has been going on behind the scenes in the Ultraleap hand tracking team. Today, we’re excited to share a Developer Preview of the fifth generation of our core software, known as Gemini.
Our tracking engine is the part of our software that turns images into digital models of your hands. For Gemini, we rewrote this from the ground up. It makes our software far more flexible for different platforms and camera hardware.
We’ve also focused on making our hand tracking even more robust. One key benefit of this is improved performance for two-handed interactions.
We know reducing friction for end users unlocks enterprise use-cases. Gemini’s improved robustness means you can easily use both hands together.— Ultraleap Developers (@ultraleap_devs) January 27, 2021
It enables more natural and immersive interaction with virtual objects and new wearable menus 😎 pic.twitter.com/bfpAdd4Wz8
Gemini’s feature improvements include:
- Even better smoothness, pose fidelity, and robustness (likely to be most apparent on desktop mode)
- Improved hand initialization
- Significantly better performance with two-hand interactions
The Gemini Developer Preview will only run on Windows 10. Releases further down the line will extend support to other platforms.
Of course, you can run Gemini on our existing camera modules (the Leap Motion Controller and Stereo IR 170). These can be mounted onto your favourite existing HMD using our VR Developer Mount or 3D-printable Stereo IR 170 Evaluation Kit mount. It allows our developer community to explore and develop applications now that will be supported in upcoming headsets.
If hands don’t appear reliably, it’s disorientating and reduces embodiment. That’s why Gemini has improved hand initialization.— Ultraleap Developers (@ultraleap_devs) January 27, 2021
Add the Stereo IR 170’s wider field of view and hands will almost certainly be tracked before they come into your sight 🙌 pic.twitter.com/vkZMrnRrlG
Robust, accurate hand tracking is one of the biggest challenges in 21st century computing. The ten years of ongoing developer support, experimentation, and feedback have been the crucible in which we’ve refined our hand tracking.
We’d love you to join us on the next stage of the journey and get involved in helping us make hand tracking work even better for your projects and applications.