# US 2022/0191562 A1 · No reference realtime video quality assessment

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

- Number: US 2022/0191562 A1
- Document ID: US20220191562A1
- Status: Published application
- Kind: A1
- Application number: 17/124,370
- Filed: 2020-12-16
- Published: 2022-06-16
- Earliest claimed: 2020-12-16
- Assignees: iStreamPlanet Co., LLC
- Applicants: iStreamPlanet Co., LLC
- Inventors: Nachiketa MISHRA; Steven ROEDER
- Web page: https://patentgrep.com/patent/US20220191562A1

## CPC

- G06V 10/77: Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- H04N 21/23418: {involving operations for analysing video streams, e.g. detecting features or characteristics (television picture signal circuitry for scene change detection H04N5/147; filtering for image enhancement G06T5/00; methods or arrangements for recognising scenes G06V20/00; arrangements characterised by components specially adapted for monitoring, identification or recognition of video in broadcast systems H04H60/59)}
- G06F 18/211: Selection of the most significant subset of features
- G06N 3/096: Transfer learning
- G06N 20/20: Ensemble learning
- G06V 10/40: Extraction of image or video features
- H04N 17/004: {for digital television systems}
- H04N 21/251: {Learning process for intelligent management, e.g. learning user preferences for recommending movies (details of learning user preferences for the retrieval of video data in a video database G06F16/739; computer systems using learning methods G06N3/08)}
- H04N 21/26603: {for automatically generating descriptors from content, e.g. when it is not made available by its provider, using content analysis techniques}
- G06N 3/045: Combinations of networks
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A method for video quality assessment implemented by a computing device, where the method includes selecting a first set of frames from an input video source, determining features of the first set of frames for input into a first machine learning model, applying the first machine learning model to the features of the first set of frames to obtain a video quality score and a confidence score, and triggering a retraining of the first machine learning model in response to the confidence score being determined to be out of bounds.

## Claims

### Claim 1

A method for video quality assessment implemented by a computing device, the method comprising:
selecting a first set of frames from an input video source;
determining features of the first set of frames for input into a first machine learning model;
applying the first machine learning model to the features of the first set of frames to obtain a video quality score and a confidence score; and
triggering a retraining of the first machine learning model in response to the confidence score being determined to be out of bounds.

### Claim 2 (depends on 1)

The method of claim 1, wherein the features include any one or more of quantization parameters, motion vectors, bit rate, scene statistics, standard deviation of quantization parameters, or standard deviation of motion vector.

### Claim 3 (depends on 1)

The method of claim 1, wherein the video quality score is generated without reference to an original frame of the input video source.

### Claim 4 (depends on 1)

The method of claim 1, further comprising:
determining the features from metadata of the first set of frames.

### Claim 5 (depends on 1)

The method of claim 1, wherein the retraining further comprises:
selecting a second set of frames of the input video source;
determining input features for a second machine learning model for the second set of frames;
applying the second machine learning model to obtain a second video quality score; and
training a third machine learning model using the second video quality score and the input features for the second machine learning model.

### Claim 6 (depends on 5)

The method of claim 5, wherein the first machine learning model is trained via transfer learning from the third machine learning model.

### Claim 7

A non-transitory machine-readable storage medium that provides instructions that, if executed by a processor, will cause said processor to perform operations comprising:
selecting a first set of frames from an input video source;
determining features of the first set of frames for input into a first machine learning model;
applying the first machine learning model to the features of the first set of frames to obtain a video quality score and a confidence score; and
triggering a retraining of the machine learning model in response to the confidence score being determined to be out of bounds.

### Claim 8 (depends on 7)

The non-transitory machine-readable storage medium of claim 7, having further instructions stored therein, which when executed by the processor cause the processor to perform additional operations, wherein the features include any one or more of quantization parameters, motion vectors, bit rate, scene statistics, standard deviation of quantization parameters, or standard deviation of motion vector.

### Claim 9 (depends on 7)

The non-transitory machine-readable storage medium of claim 7, having further instructions stored therein, which when executed by the processor cause the processor to perform additional operations, wherein the video quality score is generated without reference to an original frame of the input video source.

### Claim 10 (depends on 7)

The non-transitory machine-readable storage medium of claim 7, having further instructions stored therein, which when executed by the processor cause the processor to perform additional operations, further comprising:
determining the features from metadata of the set of frames.

### Claim 11 (depends on 7)

The non-transitory machine-readable storage medium of claim 7, having further instructions stored therein, which when executed by the processor cause the processor to perform additional operations, wherein the retraining further comprises:
selecting a second set of frames of the input video source;
determining input features for a second machine learning model for the second set of frames;
applying the second machine learning model to obtain a second video quality score; and
training a third machine learning models using the second video quality score and the input features for the second machine learning model.

### Claim 12 (depends on 11)

The non-transitory machine-readable storage medium of claim 11, having further instructions stored therein, which when executed by the processor cause the processor to perform additional operations, wherein the first machine learning model is trained via transfer learning from the third machine learning model.

### Claim 13

An electronic device comprising:
a non-transitory machine-readable storage medium having stored therein a video quality assessor; and
a processor coupled to the non-transitory machine-readable storage medium, the processor to execute the video quality assessor, the video quality assessor to select a first set of frames from an input video source, to determine features of the first set of frames for input into a first machine learning model, to apply the machine learning model to the features of the first set of frame to obtain a video quality score and a confidence score, and to trigger a retraining of the first machine learning model in response to the confidence score being determined to be out of bounds.

### Claim 14 (depends on 13)

The electronic device of claim 13, wherein the features include any one or more of quantization parameters, motion vectors, bit rate, scene statistics, standard deviation of quantization parameters, or standard deviation of motion vector.

### Claim 15 (depends on 13)

The electronic device of claim 13, wherein the video quality score is generated without reference to an original frame of the input video source.

### Claim 16 (depends on 13)

The electronic device of claim 13, the video quality assessor to further to determine the features from metadata of the first set of frames.

### Claim 17 (depends on 13)

The electronic device of claim 13, wherein the retraining further includes selecting a second set of frames of the input video source, determining input features for a second machine learning model for the second set of frames, applying the second machine learning model to obtain a second video quality score, and training a third machine learning model using the second video quality score and the input features for the second machine learning model.

### Claim 18 (depends on 17)

The electronic device of claim 17, wherein the first machine learning model is trained via transfer learning from the third machine learning model.
