Rendering-based landmark localization from 3D range images

• US 8,682,041 B2
• Filed: 01/28/2011
• Issued: 03/25/2014
• Est. Priority Date: 01/28/2011
• Status: Active Grant

• Current Assignee: Honeywell International Inc.
• Original Assignee: Honeywell International Inc.
• Inventors: McCloskey, Scott
• Primary Examiner(s): Desire, Gregory M
• Application Number:
  US13/016,448
• Publication Number:
  US 20120194504A1
• Time in Patent Office: 1,152 Days
• Field of Search: 345/419, 345/420, 382/118, 382/154, 382/203, 382/294
• US Class Current: 382/118
• CPC Class Codes: G06V 40/168 Feature extraction; Face representation

Abstract

A computer implemented method comprising acquiring a three-dimensional (3D) range image containing one or more 3D landmarks; rendering the 3D range image into a rendered visible image containing intensity patterns related to the one or more 3D landmarks; and localizing at least one of the one or more 3D landmarks in the rendered visible image to provide at least one localized 3D landmark. is provided. Such methods, and related systems, and devices, are useful in biometric identification applications.

Claims

1. A computer implemented method comprising:

acquiring a three-dimensional (3D) range image containing one or more 3D landmarks;

rendering the 3D range image into a rendered visible image containing intensity patterns related to the one or more 3D landmarks, wherein rendering the 3D range image includes generating a two-dimensional synthetic image from a scene file, and wherein the scene file includes objects in a defined language as a description of a virtual scene; and

localizing at least one of the one or more 3D landmarks in the rendered visible image to provide at least one localized 3D landmark, wherein the at least one localized 3D landmark is compared with at least one 3D landmark in a reference 3D image to provide a general identification or a biometric identification.

2. The method of claim 1 further comprising cropping the at least one localized landmark to produce an at least one cropped localized 3D landmark.

3. The method of claim 2 further comprising comparing at least one cropped localized 3D landmark with the at least one 3D landmark in the reference image to provide a general identification or a biometric identification.

4. The method of claim 1 wherein the rendered visible image is a rendered visible RGB image.

5. The method of claim 1 wherein the rendered visible image is a rendered visible black and white image or a rendered visible grayscale image.

6. The method of claim 1 wherein the one or more 3D landmarks include facial landmarks.

7. The method of claim 6 wherein the facial landmarks are selected from eyes, nose, cheekbones, chin and ear.

8. The method of claim 1 wherein the 3D range image has a subject and the rendering step comprises utilizing the 3D range image and a bidirectional reflectance distribution function (BRDF) of the subject to render the subject'"'"'s appearance relative to a synthetic light source.

9. A computer readable device having instructions to cause a computer system to implement a method comprising:

rendering a 3D range image containing one or more 3D landmarks into a rendered visible image containing intensity patterns related to the one or more 3D landmarks, wherein rendering the 3D range image includes generating a two-dimensional synthetic image from a scene file, and wherein the scene file includes objects in a defined language as a description of a virtual scene;

localizing at least one of the one or more 3D landmarks to provide at least one localized 3D landmark, wherein the at least one localized 3D landmark is compared with at least one 3D landmark in a reference 3D image to provide a general identification or a biometric identification.

10. The device of claim 9 further comprising cropping the at least one 3D landmark to produce an at least one cropped localized 3D landmark.

11. The device of claim 10 further comprising outputting the at least one localized intensity pattern to a display.

12. The device of claim 10 further comprising comparing at least one cropped localized 3D landmark with the at least one 3D landmark in the reference 3D landmark to provide a general identification or a biometric identification.

13. A system comprising:

a three-dimensional (3D) range image acquiring device associated with a 3D range image containing one or more 3D landmarks;

a rendering device configured for rendering the 3D range image into a rendered visible image containing intensity patterns related to one or more 3D landmarks, wherein rendering the 3D range image includes generating a two-dimensional synthetic image from a scene file, and wherein the scene file includes objects in a defined language as a description of a virtual scene; and

a landmark localization device in communication with the rendering device, the landmark localization device configured to localize at least one of the one or more 3D landmarks in the rendered visible image to provide at least one localized 3D landmark, wherein the at least one localized 3D landmark is compared with at least one 3D landmark in a reference 3D image to provide a general identification or a biometric identification.

14. The system of claim 13 further comprising:

a cropping device in communication with the landmark localization device to crop the at least one localized 3D landmark to produce at least one 3D landmark; and

a display in communication with the 3D range image acquiring device to output the at least one localized intensity pattern.

15. The system of claim 14 further comprising a comparing device in communication with the cropping device to compare the at least one cropped localized intensity pattern in the rendered visible image with the at least one 3D landmark in the reference 3D image wherein general identification or biometric identification can be performed.

16. The system of claim 13 wherein the 3D image acquiring device is a Coded Light Approach (CLA) device.

17. The system of claim 13 wherein the 3D acquiring device can provide pixel values of data in the 3D range image in physical distance units.

18. The system of claim 13 wherein the rendering device is a ray tracing device.

19. The system of claim 18 wherein the landmark localization device is also configured to crop the one or more 3D landmarks.

20. The system of claim 13 wherein the rendered visible image is a rendered visible RGB image.

21. The system of claim 13 wherein the rendered visible image is a rendered visible black and white image or a rendered visible grayscale image.

22. The system of claim 13 wherein the one or more 3D landmarks include facial landmarks.

23. The system of claim 22 wherein the facial landmarks are selected from distance between eyes, width of nose, depth of eye sockets, shape of cheekbones, chin features and ear features.