The Virtual Kinect
Resumo
The COVID-19 pandemic impacted researches that depended on making tests on patients and teams that had to be divided to avoid crowding in laboratories. Sharing equipment is no longer a simple strategy; the acquisition of extra equipment is beyond the means of many researchers. Research with Kinect V2 applied to body tracking suffers from sanitary restrictions, product discontinuation and limited access to newer sensors (like the Azure Kinect). Kinect V2 is an RGB-D sensor with many applications in health, ergonomics, sports, games and other areas. That is why a lot of research is still under development with it. Because of the applicability of Kinect V2 on research and the current acquisition limitations, Virtual Kinect (VK) was created. VK is an open-source solution that enables the programmer to code using Kinect SDK functions and test it without a Kinect V2 sensor or a previous recording. It simulates the behavior of a Kinect V2 sensor through MediaPipe’s Pose estimation, providing RGB image and joint tracking information, correlating the joints of the two devices. This correlation is possible due to the proximity of the devices’ joint estimates. The VK was made to be simple and practical, so that its use only depends on the DLL exchange and the use of an ordinary RGB camera.
