Crunchfish recently released a gesture skeleton network that can be tracked by standard camera sensors throughout the body. This also demonstrates the versatility of the XR Skeleton platform: through a brand-new image data training algorithm, new object detection and tracking can be incorporated in a short time.
As an illustration, Crunchfish XR Skeleton was released in 2020, and focused on gesture tracking and bare-hand interaction. With only a standard mobile phone camera, this system can detect and track 21 key points of each hand. The main target use cases of the product include AR/VR devices and smart phones, but its versatility is also applicable to other aspects, including bare-hand interaction in car and aircraft entertainment systems, and to replace TV remote controls.
Below is a demonstration video of Crunchfish XR gesture interaction:
For the skeleton model released a few days ago, the team has used Crunchfish FB Skeleton to extend pure gesture tracking to full body tracking, and it is also achieved through a standard RGB camera.
It is reported that by providing a combination of synthetic data and annotated images of the human body to the neural network, the developers created a solution capable of whole-body tracking in just a few months. FB Skeleton will be available as a commercial solution later this year.
Crunchfish Gesture Interaction R&D Director Daniel Milesson (Daniel Milesson) said: "The development speed of creating various new solutions proves the technical strength and versatility of our skeleton gesture control system."
© 2020 www.ourvrworld.com