Object identification through stereo association

摘要

Various embodiments enable a primary user to be identified and tracked using stereo association and multiple tracking algorithms. For example, a face detection algorithm can be run on each image captured by a respective camera independently. Stereo association can be performed to match faces between cameras. If the faces are matched and a primary user is determined, a face pair is created and used as the first data point in memory for initializing object tracking. Further, features of a user's face can be extracted and the change in position of these features between images can determine what tracking method will be used for that particular frame.

主权项

1. A computing device, comprising:
at least one processor; a right camera having a first field of view; a left camera having a second field of view at least partially overlapping the first field of view; memory including instructions that, when executed by the at least one processor, cause the computing device to:
capture right image data with the right camera;capture left image data with the left camera;detect a first face right representation of a first face and a second face right representation of a second face in the right image data;associate the first face right representation with a first face right bounding box and the second face right representation with a second face right bounding box;detect a third face left representation of a third face;associate the third face left representation with a third face left bounding box;compare a third face left intensity histogram for left image data of the third face left bounding box to a first face right intensity histogram for right image data of the first face right bounding box;compare the third face left intensity histogram for the left image data of the third face left bounding box to a second face right intensity histogram for right image data of the second face right bounding box;determine the first face right intensity histogram matches the third face left intensity histogram within an allowable deviation;extract, from the second face right bounding box for the second face, a set of feature points to generate a template of the second face;detect, by analyzing the left image data, a fourth face left representation of a fourth using the template of the second face; andassociate the fourth face left representation of the fourth face with a fourth face left bounding box.