# US 2022/0036535 A1 · Image quality assessment apparatus and image quality assessment method thereof

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

- Number: US 2022/0036535 A1
- Document ID: US20220036535A1
- Status: Published application
- Kind: A1
- Application number: 17/022,086
- Filed: 2020-09-16
- Published: 2022-02-03
- Earliest claimed: 2020-08-03
- Assignees: Wistron Corporation
- Applicants: Wistron Corporation
- Inventors: Yuan-Bin Lee; Hsiao-Wen Tin
- Web page: https://patentgrep.com/patent/US20220036535A1

## CPC

- G06T 7/0002: {Inspection of images, e.g. flaw detection}
- H04N 19/103: Selection of coding mode or of prediction mode
- H04N 19/136: Incoming video signal characteristics or properties
- H04N 19/154: Measured or subjectively estimated visual quality after decoding, e.g. measurement of distortion (use of rate-distortion criteria H04N19/147)
- H04N 19/176: the region being a block, e.g. a macroblock
- G06T 2207/10016: Video; Image sequence
- G06T 2207/20081: Training; Learning
- G06T 2207/20084: Artificial neural networks [ANN]
- G06T 2207/30168: Image quality inspection
- G06T 2207/30201: Face
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION

## Abstract

An image quality assessment apparatus and an image quality assessment method are provided. In the method, multiple to-be-assessed blocks are formed for an image in a video stream, these to-be-assessed blocks are inputted to a quality assessment model, and a quality of the image is determined according to an output result of the quality assessment model. The quality assessment model is trained based on a machine learning algorithm.

## Claims

### Claim 1

An image quality assessment method, comprising:
forming a plurality of to-be-assessed blocks for an image;
inputting the to-be-assessed blocks to a quality assessment model, wherein the quality assessment model is trained based on a machine learning algorithm; and
determining a quality of the image according to an output result of the quality assessment model.

### Claim 2 (depends on 1)

The image quality assessment method of claim 1, wherein the output result of quality assessment model comprises one of a plurality of objective functions, and the step of determining the quality of the image according to the output result of the quality assessment model comprises:
determining a calculated value of the objective function corresponding to each of the to-be-assessed blocks; and
determining a quality corresponding to the to-be-assessed block according to the calculated value,
wherein the output result of the quality assessment model further comprises at least one of an image feature, a perception assessment score and a statistical distribution of the to-be-assessed block.

### Claim 3 (depends on 2)

The image quality assessment method of claim 2, wherein the statistical distribution is correlated with a natural image statistical distribution, and the step of determining the calculated value of the objective function corresponding to each of the to-be-assessed blocks comprises:
determining a calculated value corresponding to each of the to-be-assessed blocks according to the statistical distribution.

### Claim 4 (depends on 2)

The image quality assessment method of claim 2, wherein the perception assessment score is obtained based on an actual rating result of human perception, a mean opinion score (MOS) database or a differential mean opinion score (DMOS) database, and the step of determining the calculated value of the objective function corresponding to each of the to-be-assessed blocks comprises:
determining a calculated value corresponding to each of the to-be-assessed blocks according to the perception assessment score.

### Claim 5 (depends on 2)

The image quality assessment method of claim 2, wherein each of the objective functions has an upper limit and a lower limit, and the step of determining the quality corresponding to the to-be-assessed block according to the calculated value comprises:
determining whether the calculated value is between the upper limit and the lower limit of the corresponding objective function.

### Claim 6 (depends on 5)

The image quality assessment method of claim 5, wherein after determining whether the calculated value is between the upper limit and the lower limit of the corresponding objective function, the method further comprises:
in response to the calculated value being between the upper limit and the lower limit, playing the image; and
in response to the calculated value not being between the upper limit and the lower limit or not having a solution, not playing the image.

### Claim 7 (depends on 1)

The image quality assessment method of claim 1, wherein the step of determining the quality of the image according to the output result of the quality assessment model comprises:
determining an encoding level of each of the to-be-assessed blocks, wherein the different encoding levels correspond to different weights; and
determining the quality of the image according to qualities of the to-be-assessed blocks and the corresponding weights.

### Claim 8 (depends on 1)

The image quality assessment method of claim 1, further comprising:
using a plurality of to-be-learned blocks in at least one scene image, a mean opinion score image database or a differential mean opinion score image database as learning data; and
using the learning data as a training sample for training the quality assessment model, wherein
perception assessment scores of a subjective assessment of the to-be-learned blocks are counted; and
the perception assessment scores corresponding to the to-be-learned blocks are used as the training sample for training the quality assessment model.

### Claim 9 (depends on 1)

The image quality assessment method of claim 1, further comprising:
determining one of a plurality of objective functions according to at least one of an image feature, a perception assessment score and a statistical distribution corresponding to a plurality of to-be-learned blocks, and using the plurality of to-be-learned blocks and the objective function as a training sample for training the quality assessment model.

### Claim 10 (depends on 1)

The image quality assessment method of claim 1, further comprising:
re-encoding the image according to the quality of the image.

### Claim 11

An image quality assessment apparatus, comprising:
a communication transceiver, configured to receive an image; and
a processor, coupled to the communication transceiver, and configured to:
form a plurality of to-be-assessed blocks for the image;
input the to-be-assessed blocks to a quality assessment model, wherein the quality assessment model is trained based on a machine learning algorithm; and
determine a quality of the image according to an output result of the quality assessment model.

### Claim 12 (depends on 11)

The image quality assessment apparatus of claim 11, wherein the output result of quality assessment model comprises one of a plurality of objective functions, and the processor is further configured to:
determine a calculated value of the objective function corresponding to each of the to-be-assessed blocks; and
determine a quality corresponding to the to-be-assessed block according to the calculated value,
wherein the output result of the quality assessment model further comprises at least one of an image feature, a perception assessment score and a statistical distribution of the to-be-assessed block.

### Claim 13 (depends on 12)

The image quality assessment apparatus of claim 12, wherein the statistical distribution is correlated with a natural image statistical distribution, and the processor is further configured to:
determine a calculated value corresponding to each of the to-be-assessed blocks according to the statistical distribution.

### Claim 14 (depends on 12)

The image quality assessment apparatus of claim 12, wherein the perception assessment score is obtained based on an actual rating result of human perception, a mean opinion score database or a differential mean opinion score database, and the processor is further configured to:
determine a calculated value corresponding to each of the to-be-assessed blocks according to the perception assessment score.

### Claim 15 (depends on 12)

The image quality assessment apparatus of claim 12, wherein each of the objective functions has an upper limit and a lower limit, and the processor is further configured to:
determine whether the calculated value is between the upper limit and the lower limit of the corresponding objective function.

### Claim 16 (depends on 15)

The image quality assessment apparatus of claim 15, further comprising:
a display, coupled to the processor, wherein the processor is configured to:
in response to the calculated value being between the upper limit and the lower limit, play the image through the display; and
in response to the calculated value not being between the upper limit and the lower limit or not having a solution, not play the image through the display.

### Claim 17 (depends on 11)

The image quality assessment apparatus of claim 11, wherein the processor is further configured to:
determine an encoding level of each of the to-be-assessed blocks, wherein the different encoding levels correspond to different weights; and
determine the quality of the image according to qualities of the to-be-assessed blocks and the corresponding weights.

### Claim 18 (depends on 11)

The image quality assessment apparatus of claim 11, wherein the processor is further configured to:
use a plurality of to-be-learned blocks in at least one scene image, a mean opinion score image database or a differential mean opinion score image database as learning data; and
use the learning data as a training sample for training the quality assessment model, wherein
perception assessment scores of a subjective assessment of the to-be-learned blocks are counted; and
the perception assessment scores corresponding to the to-be-learned blocks are used as the training sample for training the quality assessment model.

### Claim 19 (depends on 11)

The image quality assessment apparatus of claim 11, wherein the processor is further configured to:
determine one of a plurality of objective functions according to at least one of an image feature, a perception assessment score and a statistical distribution corresponding to a plurality of to-be-learned blocks, and using the plurality of to-be-learned blocks and the objective function as a training sample for training the quality assessment model.

### Claim 20 (depends on 11)

The image quality assessment apparatus of claim 11, wherein the processor is further configured to:
send a request for re-encoding the image through the communication transceiver according to the quality of the image.
