Method for pre-processing an image in facial recognition system

• US 8,798,391 B2
• Filed: 09/23/2011
• Issued: 08/05/2014
• Est. Priority Date: 05/23/2011
• Status: Active Grant

• Current Assignee: Infosys Limited
• Original Assignee: Infosys Limited
• Inventors: Tallamraju, Ravindra Babu, Kalmanje, Mahalakshmi, Sreenivasan Venkateshaiah, Subrahmanya, Danivas, Chethan Shikaripur Annajirao, Chandrappa, Pavitra
• Primary Examiner(s): Alavi, Amir
• Application Number:
  US13/241,786
• Publication Number:
  US 20120301020A1
• Time in Patent Office: 1,047 Days
• Field of Search: 382/154, 382/162, 382/164, 382/165, 382/167, 382/173, 382/190, 382/224, 358/515, 358/518, 358/520, 358/530, 345/49, 345/150, 345/199, 348/586, 348589-593
• US Class Current: 382/268
• CPC Class Codes: G06V 10/44 Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
• CPC Class Codes: G06V 40/162 using pixel segmentation or colour matching

Abstract

A method preprocessing scheme that improves classification accuracy of a face recognition system comprises identifying a plurality of edges of an image. The method further comprises step of identifying a plurality of discontinuities in the plurality of edges of the image. Furthermore, the method also comprises generating contiguous outline for the image. The last step of the method comprises filling a background of the image with at least one color.

Claims

1. A method for improving classification accuracy of a face recognition system, said method comprising

identifying a plurality of edges of an image,

identifying a plurality of discontinuities in one or more edges of said identified plurality of edges of said image, wherein said discontinuities in said one or more edges correspond to missing edge pixels,

computing said missing edge pixels, and

generating a contiguous outline for said image using said computed pixels and said identified plurality of edges,

from said contiguous outline, computing a set of pixels that complements said image, and

filling said computed set of pixels with gray to maximize classification accuracy of said face recognition system.

2. The method of claim 1, further comprising generating said plurality of edges to said image using the Sobel operator.

3. The method of claim 1, further comprising receiving said image.

4. The method of claim 1, wherein identifying said plurality of discontinuities in said plurality of edges of said image further comprises preventing said gray from being applied to the entire image.

5. The method of claim 1, wherein said contiguous outline for said image is generated using linear interpolation.

6. The method of claim 1, further comprising isolating a face portion of said image.

7. The method of claim 6, further comprising applying at least one color to said image, wherein a value of said at least one color is in a range of about 0 to 255.

8. A computer program product a non-transitory computer-readable medium having instructions encoded thereon for enabling a processor to perform the operations of

identifying a plurality of edges of an image,

identifying a plurality of discontinuities in one or more edges of said identified plurality of edges of said image, wherein said discontinuities in said one or more edges correspond to missing edge pixels,

computing said missing edge pixels and generating a contiguous outline for said image using said computed pixels and said identified plurality of edges,

from said contiguous outline, computing a set of pixels that complements said image, and

filling said computed set of pixels with gray to maximize classification accuracy of said face recognition system.

9. The computer program product of claim 8, wherein said instructions further comprise instructions for generating said plurality of edges to said image using the Sobel operator.

10. The computer program product of claim 8, wherein said instructions further comprise instructions for receiving said image.

11. The computer program product of claim 8, wherein said instructions further comprise instructions for preventing gray from being applied on said entire image.

12. The computer program product of claim 8, wherein said instructions further comprise instructions for generating said contiguous outline for said image using linear interpolation.

13. The computer program product of claim 8, wherein said instructions further comprise instructions for isolating a face portion of said image.