# US 2022/0036527 A1 · Method and image processing device for mura detection on display

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2022/0036527 A1
- Document ID: US20220036527A1
- Status: Published application
- Kind: A1
- Application number: 16/944,165
- Filed: 2020-07-31
- Published: 2022-02-03
- Earliest claimed: 2020-07-31
- Assignees: Novatek Microelectronics Corp.
- Applicants: Novatek Microelectronics Corp.
- Inventors: Chih-Yu Chu; Po-Yuan Hsieh; Chieh-En Lee; Chung-Hao Tien; Shih-Hsuan Chen
- Web page: https://patentgrep.com/patent/US20220036527A1

## CPC

- G06T 7/001: {using an image reference approach}
- G06F 18/22: Matching criteria, e.g. proximity measures
- G06T 5/20: using local operators
- G06T 7/0002: {Inspection of images, e.g. flaw detection}
- G06T 7/11: Region-based segmentation
- G06V 10/761: {Proximity, similarity or dissimilarity measures}
- G06T 2207/20021: Dividing image into blocks, subimages or windows
- G06T 2207/30168: Image quality inspection
- G06V 10/25: Determination of region of interest [ROI] or a volume of interest [VOI]
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

A method and an image processing device for mura detection on a display are proposed. The method includes the following steps. An original image of the display is received and segmented into region of interest (ROI) patches. A predetermined range of spatial frequency components are filtered out from the ROI patches to generate filtered ROI patches. A mura defect is identified from the display according to the filtered ROI patches and predetermined mura patterns.

## Claims

### Claim 1

A method for mura detection on a display, comprising:
receiving an original image of the display;
segmenting the original image into a plurality of region of interest (ROI) patches;
filtering out a predetermined range of spatial frequency components from the ROI patches to generate a plurality of filtered ROI patches; and
identifying a mura defect from the display according to the filtered ROI patches and a plurality of predetermined mura patterns.

### Claim 2 (depends on 1)

The method according to claim 1, wherein the step of filtering out the predetermined range of the spatial frequency components from the ROI patches to generate the filtered ROI patches comprises:
transforming the ROI patches from a spatial domain into a spatial frequency domain to generate a plurality of transformed ROI patches;
filtering out the predetermined range of the spatial frequency components from the transformed ROI patches to generate a plurality of filtered-transformed ROI patches; and
inverse-transforming the filtered-transformed ROI patches from the spatial frequency domain to the spatial domain to generate the filtered ROI patches.

### Claim 3 (depends on 1)

The method according to claim 1, wherein the step of filtering out the predetermined range of the spatial frequency components from the ROI patches comprises:
filtering out the predetermined range of the spatial frequency components from the ROI patches according to a contrast sensitivity function.

### Claim 4 (depends on 3)

The method according to claim 3, wherein the contrast sensitivity function is associated with a human's visual perception.

### Claim 5 (depends on 1)

The method according to claim 1, wherein the predetermined range of the spatial frequency components comprises a range corresponding to a human's low visual perception.

### Claim 6 (depends on 1)

The method according to claim 1, wherein each of the predetermined mura patterns corresponds to a visibility threshold, and the step of identifying the mura defect from the display according to the filtered ROI patches and the predetermined mura patterns comprises:
for each of the filtered ROI patches, determining one of the visibility thresholds corresponding to the predetermined mura patterns to be a matched visibility threshold of the filtered ROI patch by performing similarity comparison based on the filtered ROI patch and the predetermined mura patterns; and
identifying the mura defect from the display according to the filtered ROI patches and the matched visibility thresholds of all the filtered ROI patches.

### Claim 7 (depends on 6)

The method according to claim 6, wherein for each of the filtered ROI patches, the step of determining one of the visibility threshold to be the matched visibility threshold of the filtered ROI patch by performing similarity comparison based on the filtered ROI patch and the predetermined mura patterns comprises:
determining a contrast ROI patch corresponding to the filtered ROI patch; and
determining a matched mura pattern with a highest similarity to the contrast ROI patch from the predetermined mura patterns to obtain the visibility threshold corresponding to the matched mura pattern as the matched visibility threshold of the filtered ROI patch.

### Claim 8 (depends on 7)

The method according to claim 7, wherein for each of the filtered ROI patches, the step of determining the contrast ROI patch corresponding to the filtered ROI patch comprises:
performing background estimation on the filtered ROI patch to generate a background ROI patch; and
computing the contrast ROI patch corresponding to the filtered ROI patch according to the filtered ROI patch and the background ROI patch.

### Claim 9 (depends on 8)

The method according to claim 8, wherein for each of the filtered ROI patches, the step of performing background estimation on the filtered ROI patch to generate the background ROI patch comprises:
performing low-pass filtering on the filtered ROI patch to generate the background ROI patch.

### Claim 10 (depends on 8)

The method according to claim 8, wherein for each of the filtered ROI patches, the step of computing the contrast ROI patch corresponding to the filtered ROI patch according to the filtered ROI patch and the background ROI patch comprises:
evaluating a division of the filtered ROI patch and the background ROI patch to obtain the contrast ROI patch.

### Claim 11 (depends on 6)

The method according to claim 6, wherein the step of identifying the mura defect from the display according to the contrast ROI patches and the visibility thresholds of all the filtered ROI patches comprises:
for each of the filtered ROI patches, comparing the contrast ROI patches and the visibility threshold to obtain a mura image comprising a plurality of mura levels of the filtered ROI patch; and
identifying the mura defect from the display according to the mura image.

### Claim 12 (depends on 11)

The method according to claim 11, wherein each of the mura levels corresponds to a different just noticeable difference (JND) threshold.

### Claim 13 (depends on 11)

The method according to claim 11, wherein the mura levels comprise a first mura level taking a positive value and a second mura level taking a negative value, wherein a brightness of a first pixel corresponding to the first mura level is greater than an average brightness of a region of the display corresponding to the filtered ROI patch, and wherein a brightness of a second pixel corresponding to the second mura level is less than the average brightness of the region of the display corresponding to the filtered ROI patch.

### Claim 14 (depends on 11)

The method according to claim 11, wherein a larger magnitude of a first mura level among the mura levels corresponds to a larger severity of the mura defect.

### Claim 15 (depends on 1)

The method according to claim 1, wherein after the step of segmenting the original image into the ROI patches comprises:
segmenting the original image with a plurality of different sizes to respectively generate the ROI patches, wherein each of the sizes is associated with a different view angle.

### Claim 16 (depends on 15)

The method according to claim 15, wherein each of the predetermined mura patterns corresponds to a visibility threshold, and the step of identifying the mura defect from the display according to the filtered ROI patches and the predetermined mura patterns comprises:
for each of the filtered ROI patches in each of the different sizes, determining one of the visibility thresholds corresponding to the predetermined mara patterns to be a matched visibility threshold of the filtered ROI patch by performing similarity comparison based on the filtered ROI patch and the predetermined mura patterns; and
identifying the mura defect from the display according to the contrast ROI patches and the matched visibility thresholds of all the filtered ROI patches for all the sizes.

### Claim 17 (depends on 16)

The method according to claim 16, the step of identifying the mura defect from the display according to the contrast ROI patches and the matched visibility thresholds of all the filtered ROI patches for all the sizes comprises:
for each of the different sizes:
for each of the filtered ROI patches, comparing the contrast ROI patches and the matched visibility threshold to obtain a mura image comprising a plurality of mura levels of the filtered ROI patch; and
computing a weighted sum of the mura images respectively corresponding to the different sizes to obtain a weighted mura image; and
identifying the mura defect from the display according to the weighed mura image.

### Claim 18 (depends on 1)

The method according to claim 1, wherein the display is an OLED display.

### Claim 19

A method for mura detection on a display, comprising:
receiving an original image of the display;
segmenting the original image into a plurality of region of interest (ROI) patches; and
filtering out a predetermined range of spatial frequency components from the ROI patches to generate a plurality of filtered ROI patches according to a human's visual perception.

### Claim 20 (depends on 19)

The method according to claim 19, wherein the predetermined range of the spatial frequency components comprises a range corresponding to a human's low visual perception.

### Claim 21

An image processing device comprising:
a memory circuit, configured to store a plurality of predetermined mura patterns;
a processing circuit, configured to:
receive an original image of a display;
segment the original image into a plurality of region of interest (ROI) patches;
filter out a predetermined range of spatial frequency components from the ROI patches to generate a plurality of filtered ROI patches; and
identify a mura defect from the display according to the filtered ROI patches and the predetermined mura patterns.
