Deep image de-aliasing technique

Abstract

Techniques for anti-aliasing depth images are provided. A depth image may be generated based on the phase difference between the transmitted and received modulated light beams. The method includes accessing a depth image having depth values ​​therein for a plurality of locations therein. Each location has one or more neighbor locations. A potential depth value is determined for each of the plurality of locations based on a depth value for the location in the depth image and potential aliasing in the depth image. A cost function is determined based on the difference between the potential depth values ​​of each location and its neighbors. Determining the cost function includes assigning a higher cost to larger differences in potential depth values ​​between adjacent locations. This cost function is sufficiently minimized to select a potential depth value for each location.

Description

technical field

[0001] The present invention relates to image processing, and in particular to anti-aliasing of depth images. Background technique

[0002] A depth camera system obtains data about the location of a person or other object within a physical space. This information may be referred to as "depth information". Depth information can be input to applications within the computing system for a variety of applications. There may be many applications such as for military, entertainment, sports and medical purposes. For example, depth information about a person can be mapped to a three-dimensional (3-D) human skeletal model and used to create an animated character or avatar.

[0003] To determine depth information, a depth camera may project light onto objects in the camera's field of view. The light reflects off the subject and returns to the camera where it is processed to determine depth information. One technique for determining distance to an object is based on...

Images