System and method for mura detection on a display

Patent metadata

Field	Value
Publication number	US-10681344-B2
Application number	US-201815909893-A
Country	US
Kind code	B2
Filing date	Mar 1, 2018
Priority date	Dec 15, 2017
Publication date	Jun 9, 2020
Grant date	Jun 9, 2020

Abstract

A system and method for white spot Mura defects on a display. The system is configured to pre-process an input images to generate a plurality of image patches. A feature vector is then extracted for each of the plurality of image patches. The feature vector includes at least one image moment feature and at least one texture feature. A machine learning classifier then determines the presence of a defect in each patch using the feature vector.

First claim

What is claimed is:

1. A system for identifying Mura in a display, the system comprising:
a memory;
a processor configured to execute instructions stored on the memory that, when executed by the processor, cause the processor to:
pre-process an input image, wherein pre-processing an input image comprises generating a plurality of image patches by identifying at least one local maxima candidate in the input image and generating an image patch for each of the at least one local maxima candidate;
extract a feature vector for each of the plurality of image patches, wherein the feature vector comprises at least one image moment feature and at least one texture feature;
classify each image patch based on a presence of a defect by providing the feature vector for each image patch to a machine learning classifier,
wherein generating a plurality of image patches further comprises:
adding each identified local maxima candidate to a candidate list; and
generating an image patch for each local maxima in the candidate list, and

wherein each image patch is centered corresponding to the corresponding local maxima candidate.

2. The system of claim 1 , wherein each image patch is centered at the corresponding local maxima candidate.

3. The system of claim 2 , wherein generating a plurality of image patches further comprises filtering local maxima candidates in the candidate list by removing each local maxima candidate from the candidate list when the local maxima candidate has a value less than a noise tolerance threshold.

4. The system of claim 3 , wherein generating a plurality of image patches further comprises:
dividing the input image into a plurality of areas;
identifying a maximum local maxima in each area of the plurality of areas; and
removing all local maxima from the local maxima list except for each maximum local maxima.

5. The system of claim 1 , wherein the machine learning classifier comprises a support vector machine.

6. The system of claim 1 , wherein the at least one texture feature comprises at least one of a correlation Gray-Level Co-Occurrence Matrix (GLCM) or a contrast GLCM.

7. The system of claim 1 , wherein the at least one image moment comprises at least one of a mu 30 moment, hu 1 moment, or a hu 5 moment.

8. The system of claim 1 , wherein pre-process an input image further comprises performing Gaussian smoothing on the input image and normalizing the smoothed input image by mapping a dynamic range of the smoothed input image to an expected range.

9. The system of claim 1 , wherein the defect comprises white spot Mura.

10. A method for identifying Mura in a display, the method comprising:
pre-processing an input image, by a processor, wherein pre-processing the input image comprises generating a plurality of image patches by identifying at least one local maxima candidate in the input image and generating an image patch for each of the at least one local maxima candidate;
extracting a feature vector, by a processor, for each of the plurality of image patches, wherein the feature vector comprises at least one image moment feature and at least one texture feature;
classifying each image patch based on a presence of a defect by providing the feature vector for each image patch to a machine learning classifier,
wherein generating a plurality of image patches further comprises:
adding each identified local maxima candidate to a candidate list; and
generating an image patch for each local maxima in the candidate list, and

wherein each image patch is centered corresponding to the corresponding local maxima candidate.

11. The system of claim 10 , wherein each image patch is centered at the corresponding local maxima candidate.

12. The system of claim 11 , wherein generating a plurality of image patches further comprises filtering local maxima candidates in the candidate list by removing each local maxima candidate from the candidate list when the local maxima candidate has a value less than a noise tolerance threshold.

13. The system of claim 12 , wherein generating a plurality of image patches further comprises:
dividing the input image into a plurality of areas;
identifying a maximum local maxima in each area of the plurality of areas; and
removing all local maxima from the local maxima list except for each maximum local maxima.

14. The system of claim 10 , wherein the machine learning classifier comprises a support vector machine.

15. The system of claim 10 , wherein the at least one texture feature comprises at least one of a correlation Gray-Level Co-Occurrence Matrix (GLCM) or a contrast GLCM.

16. The system of claim 10 , wherein the at least one image moment comprises at least one of a mu 30 moment, hu 1 moment, or a hu 5 moment.

17. The system of claim 10 , wherein pre-processing an input image further comprises performing Gaussian smoothing on the input image and normalizing the smoothed input image by mapping a dynamic range of the smoothed input image to an expected range.

18. The system of claim 10 , wherein the defect comprises white spot Mura.

19. A method for identifying Mura in a display, the method comprising:
pre-processing an input image, by a processor, wherein pre-processing the input image comprises generating a plurality of image patches, wherein generating a plurality of image patches comprises:
identifying at least one local maxima candidate in the input image;
adding each identified local maxima candidate to a candidate list; and
generating an image patch for each local maxima in the candidate list, wherein each image patch is centered corresponding to the corresponding local maxima candidate;

extracting a feature vector, by a processor, for each of the plurality of image patches, wherein the feature vector comprises at least one image moment feature and at least one texture feature;
classifying each image patch, by a processor, based on a presence of a white spot Mura defect by providing the feature vector for each image patch to a support vector machine.

20. The method of claim 19 , wherein the at least one texture feature comprises at least one of a correlation Gray-Level Co-Occurrence Matrix (GLCM) or a contrast GLCM and the at least one image moment comprises at least one of a mu 30 moment, hu 1 moment, or a hu 5 moment.

21. A white spot MURA candidate detector comprising:
a local maxima finder configured to identifying at least one local maxima candidate in an input image, the local maxima finder comprising:
a local maxima calculator configured to identify each local maxima in the input image and add each local maxima to a candidate list;
a local maxima sorter configured to sort the local maxima list by a value; and
a noise tolerance filter configured to remove any local maxima from the candidate list below a noise tolerance level; and

an image patch generator configured to generating an image patch for each local maxima in the candidate list, wherein each image patch is centered at the corresponding local maxima candidate.

Assignees

• Samsung Display Co Ltd

Inventors

• Zhang Yiwei
• Lee Janghwan

Classifications

• G09G2360/145

  the light originating from the display screen · CPC title

• G09G3/006 Primary

  Electronic inspection or testing of displays and display drivers, e.g. of LED or LCD displays (testing individual LED's G01R31/2635; testing lamps G01R31/44; testing of optical features of LCD displays G02F1/1309) · CPC title

• G06T2207/20081

  Training; Learning · CPC title

• G06T7/0004

  Industrial image inspection · CPC title

• G09G2320/0233

  Improving the luminance or brightness uniformity across the screen · CPC title