December 7, 2021

Understand Depth Sensor, Gesture Recognition – Some Interesting Videos





Found some interesting videos about Kinect, but the technology is directly related to HoloLens. Hope you enjoy.


The first video is How the “Kinect Depth Sensor Works in 2 Minutes”. The video illustrates how depth sensor works.

Two cameras from two different view points are required, plus the object to form a triangle, so you can calculate the depth of the object. The two camera approach has some limitations. Kinect uses infrared light to improve that by project a speckle pattern to code each spot, so that the RGB camera can identify an object even they are the same color. I am not sure if the sun light limitation applies to the IR sensor in HoloLens though.


The second video here is the work done by Microsoft MVP Josh Blake of InfoStrat. He set up a test on Kinect v1 and Kinect v2, found drastic improvement in response time and details. As you can see from the video, Kinect v2 detects much more skeletal details of his body (ignore the legs), especially his hands: not only hand-close, hand-open, but finger pushes, two fingers up or one finger up. I am assuming HoloLens is in the v3 class.


The third one demonstrates the use of Kinect in retail clothing scenario: try on clothes without go to the fitting room. The concept is old, but it’s value is there. I hope HoloLens can popularize this idea.


HoloLens, Recommended Readings – Business Ideas Collection