The UserExtraction™ processes the color image and the 3D depth image using innovative computer vision and image processing algorithms to identify a human face and body, perfectly delineate the edges, and eliminate the background. Nuvixa UserExtraction™ is robust to different dynamic environments. However, in order to obtain a high quality picture and user edge from
UserExtraction™, we summarize the recommended best practices below:
Position the Kinect™ so that it is directly in front with the user facing the camera, such that his/her head is within the field of view.
For the best results, position yourself three to six feet from the Kinect™ camera.
Position yourself in front of a plain and simple background. Also, wear clothing that has distinctly different colors from the background.
Use a well lit room, preferably lit by fluorescent lighting
If the camera is failing to sense part or all of a user, wave hand within 6 inches from the camera to reset UserExtraction™.
Wave at the camera when you enter the field of view of the camera so the camera senses movement recognizing a human presence
Positioning oneself within 2.5 feet of the camera: The Kinect™ 3D depth information degrades when the subject is within three feet. This can cause gaps or holes in the user.
Wearing clothing similar in color to the background: This may cause a wavy edge around the user due to difficult in perfectly delineating the user from the background.
Standing in front of a very busy or complex background: This could cause a wavy or jagged edge around the user.
Standing in front of a glass window: The Kinect™ 3D technology is unable to measure thedepth of glass windows. The lack of 3D information about the background can cause inaccurate edges.
Using poor or dim lighting
Direct exposure to the sunlight: This can interfere with the Kinect™ 3D depth quality causing inaccurate edges.
Standing on the fringes of the field of view of the camera