Use of three-dimensional top-down views for business analytics

1. A method of analyzing a depth image in a digital system, the method comprising:

detecting a foreground object in a depth image, wherein the depth image is a top-down perspective of a scene; and

performing data extraction and classification on the foreground object using depth information in the depth image;

computing a height of the foreground object based on a height of a depth camera used to capture the depth image; and

classifying the foreground object based on the computed height and a 2D shape of the foreground object.

2. The method of claim 1 , further comprising performing data extraction and classification using both XY spatial shape of the foreground object in the depth image and depth information in the depth image.

3. The method of claim 1 , wherein computing a height comprises:

determining a smallest depth value of pixels in the depth image corresponding to the foreground object; and

computing the height as a difference between the height of the depth camera and the smallest depth value.

4. The method of claim 1 , wherein performing data extraction and classification further comprises:

classifying the foreground object based on the computed height.

5. The method of claim 4 , wherein classifying the foreground object further comprises classifying the foreground object as an adult or a child based on expected heights for an adult and a child.

6. The method of claim 1 , wherein classifying the foreground object further comprises classifying the foreground object as an animal when the computed height is within an expected height range for an animal and the 2D shape of the foreground object corresponds to an expected 2D shape for an animal.

7. The method of claim 6 , further comprising determining whether a foreground object classified as a person has been identified in proximity to the foreground object if the foreground object is classified as an animal.

8. The method of claim 1 , wherein classifying the foreground object further comprises:

classifying the foreground object as a shopping cart when the computed height is an expected height for a shopping cart and the 2D shape of the foreground object corresponds to an expected 2D shape for a shopping cart; and

determining fullness of the shopping cart based on depth values inside the foreground object.

9. The method of claim 1 , wherein the depth image is a depth image captured by one selected from a group consisting of a stereoscopic camera, a time-of-flight camera, and a structured light camera.