System and method for recognizing a 3-D object by generating a rotated 2-D image of the object from a set of 2-D enrollment images

• US 6,026,188 A
• Filed: 10/10/1997
• Issued: 02/15/2000
• Est. Priority Date: 10/10/1997
• Status: Expired due to Term

• Current Assignee: Unisys Corporation
• Original Assignee: Unisys Corporation
• Inventors: Dionysian, Raffi
• Primary Examiner(s): Boudreau, Leo H.
• Assistant Examiner(s): Mariam, Daniel G.
• Application Number:
  US08/947,479
• Time in Patent Office: 858 Days
• Field of Search: 382/115, 382/118, 382/169, 382/209, 382/216, 250/334, 345/427, 348/164, 348/50, 348/143, 348/139
• US Class Current: 382/216
• CPC Class Codes: G06T 3/60 Rotation of whole images or parts thereof
• CPC Class Codes: G06V 10/7515 Shifting the patterns to accommodate for positional errors
• CPC Class Codes: G06V 40/172 Classification, e.g. identification

Abstract

A three-dimensional object X is recognized by a method which includes the steps of: a) storing in a memory, a set of two-dimensional enrollment images X₁ -X_n of the object X as viewed from n different directions; b) acquiring a two-dimensional access image of an unidentified object as viewed from a new direction which lies between the n different directions; c) generating, from the stored set of images, a rotated two-dimensional image of the object X as viewed from a direction which is closer to the new direction than any of the n different directions; and, d) identifying the unidentified object as the object X only if a correlation between the image of the unidentified object and the rotated image exceeds a predetermined threshold. By generating the rotated image, misalignments with the access image of the unidentified object are reduced; and thus the predetermined threshold can be raised to reduce recognition errors.

Claims

1. A method of electronically recognizing a three-dimensional object X;

said method including the steps of;

storing in a memory, a set of two-dimensional images X₁ -X_n of said object X as viewed from at least three different directions;

acquiring a two-dimensional image of an unidentified object as viewed from a new direction which lies between said different directions;

automatically generating, from said stored set of images, a rotated two-dimensional image of said object X as viewed from a direction which is closer to said new direction than any of said different directions; and

,

identifying said unidentified object as said object X only if a correlation between said image of said unidentified object and said rotated image exceeds a predetermined threshold.

2. A method according to claim 1 wherein said generating step includes the substeps of:

measuring distances between particular features in said image of said unidentified object, and distances between the same features in one of said images X₁ -X_n ; and

producing said rotated two-dimensional image of said object X by interpolating a subset of said images X₁ -X_n as a function of a ratio of the measured distances.

3. A method according to claim 2 wherein said object X and said unidentified object are a human face.

4. A method according to claim 3 wherein said particular features include a nose on said face, and the side of said face.

5. A method according to claim 3 wherein said particular features include a nose on said face, and the top of said face.

6. A method according to claim 3 wherein said particular features include an eye on said face, and the side of said face.

7. A method according to claim 3 wherein said particular features include an eye on said face, and the top of said face.

8. A method according to claim 3 and further including the step of enabling said unidentified object to enter a secure area only if said correlation exceeds said predetermined threshold.

9. A method according to claim 1 wherein said generating step includes the substeps of:

selecting four images in said set which have the largest correlation to said image of said unidentified object; and

producing said rotated two-dimensional image of said object X by interpolating the four selected images.

10. A method according to claim 1 wherein said generating step includes the substeps of:

selecting three images in said set which have the largest correlation to said image of said unidentified object; and

producing said rotated two-dimensional image of said object X by interpolating the three selected images.

11. A method according to claim 1 wherein said set of two-dimensional images includes a total of nine images.

12. A method according to claim 1 wherein said set of two-dimensional images includes a total of five images.

13. A method according to claim 1 wherein said directions differ from each other by less than thirty degrees.

14. A method according to claim 1 wherein said set of images of said object X and said image of said unidentified object are in the infrared frequency spectrum.

15. A method according to claim 1 wherein said set of images of said object X and said image of said unidentified object are in the visible frequency spectrum.

16. An electronic system for recognizing a three-dimensional object X;

said system being comprised of;

a memory for storing a set of two-dimensional images X₁ -X_n of said object X as viewed from at least three different directions;

a camera for acquiring a two-dimensional image of an unidentified object as viewed from a new direction which lies between said different directions;

an image rotating means for automatically generating from said stored set of images, a rotated two-dimensional image of said object X as viewed from a direction which is closer to said new direction than any of said different directions; and

,

a correlating means, for identifying said unidentified object as said object X only if a correlation between said image of said unidentified object and said rotated image exceeds a predetermined threshold.