Method for generating metadata, image processing method, electronic device, and program product

Patent metadata

Field	Value
Publication number	US-12067694-B2
Application number	US-202217672369-A
Country	US
Kind code	B2
Filing date	Feb 15, 2022
Priority date	Jan 21, 2022
Publication date	Aug 20, 2024
Grant date	Aug 20, 2024

Abstract

A method in an illustrative embodiment includes dividing a first image having a first resolution into a plurality of image patches, and converting each of the plurality of image patches into an image patch having a second resolution by using a plurality of candidate conversion models, the second resolution being higher than the first resolution. The method may include determining a plurality of quality factors corresponding to the plurality of candidate conversion models for each of the converted image patches. The method may include selecting a conversion model for each image patch from the plurality of candidate conversion models based on the plurality of quality factors and computation load factors of the plurality of candidate conversion models, and generating metadata for converting the first image into a second image having the second resolution based on position information of each image patch and the conversion model selected for each image patch.

First claim

What is claimed is:

1. A method for generating metadata, comprising:
dividing a first image having a first resolution into a plurality of image patches;
converting each of the plurality of image patches into an image patch having a second resolution by using a plurality of candidate conversion models, the second resolution being higher than the first resolution;
determining a plurality of quality factors corresponding to the plurality of candidate conversion models for each of the converted image patches;
selecting a conversion model for each image patch from the plurality of candidate conversion models based on the plurality of quality factors and computation load factors of the plurality of candidate conversion models; and
generating the metadata for converting the first image into a second image having the second resolution based on position information of each image patch and the conversion model selected for each image patch.

2. The method according to claim 1 , wherein the plurality of candidate conversion models comprise at least a first conversion model having a first computation load and a second conversion model having a second computation load, the first computation load being smaller than the second computation load.

3. The method according to claim 2 , wherein selecting the conversion model for each image patch comprises:
selecting, for each image patch, the first conversion model for the image patch responsive to determining that a difference between a second quality factor of the image patch corresponding to the second conversion model and a first quality factor of the image patch corresponding to the first conversion model is smaller than or equal to a first predetermined threshold.

4. The method according to claim 2 , wherein selecting the conversion model for each image patch comprises:
selecting, for each image patch, the first conversion model for the image patch responsive to determining that a ratio of a difference between a second quality factor of the image patch corresponding to the second conversion model and a first quality factor of the image patch corresponding to the first conversion model to the first quality factor is smaller than or equal to a second predetermined threshold.

5. The method according to claim 1 , wherein the first image is a first video comprising a plurality of image frames and the second image is a second video comprising a plurality of image frames, the method further comprising:
sending the first video and the metadata to a video requester to convert the first video into the second video.

6. The method according to claim 5 , wherein the position information indicates positions of each image patch in corresponding image frames of the first video and the second video.

7. The method according to claim 1 , further comprising:
receiving the first image having the first resolution and the metadata for image conversion, the first image comprising the plurality of image patches, and the metadata indicating a corresponding conversion model, used for converting each of the plurality of image patches, among the plurality of candidate conversion models;
respectively converting the plurality of image patches into a plurality of image patches having the second resolution based on the metadata; and
generating the second image having the second resolution based on the plurality of converted image patches.

8. The method according to claim 7 , wherein the metadata is generated based on position information of each image patch and a conversion model selected for each image patch, and a plurality of candidate conversion models comprise at least a first conversion model having a first computation load and a second conversion model having a second computation load, the first computation load being smaller than the second computation load.

9. The method according to claim 8 , wherein for each image patch,
the conversion model selected for the image patch is the first conversion model responsive to determining that a difference between a second quality factor of the image patch corresponding to the second conversion model and a first quality factor of the image patch corresponding to the first conversion model is smaller than or equal to a first predetermined threshold.

10. The method according to claim 8 , wherein for each image patch,
the conversion model selected for the image patch is the first conversion model responsive to determining that a ratio of a difference between a second quality factor of the image patch corresponding to the second conversion model and a first quality factor of the image patch corresponding to the first conversion model to the first quality factor is smaller than or equal to a second predetermined threshold.

11. An electronic device, comprising:
a processor; and
a memory coupled to the processor and having instructions stored therein which, when executed by the processor, cause the electronic device to perform actions comprising:
dividing a first image having a first resolution into a plurality of image patches;
converting each of the plurality of image patches into an image patch having a second resolution by using a plurality of candidate conversion models, the second resolution being higher than the first resolution;
determining a plurality of quality factors corresponding to the plurality of candidate conversion models for each of the converted image patches;
selecting a conversion model for each image patch from the plurality of candidate conversion models based on the plurality of quality factors and computation load factors of the plurality of candidate conversion models; and
generating metadata for converting the first image into a second image having the second resolution based on position information of each image patch and the conversion model selected for each image patch.

12. The electronic device according to claim 11 , wherein the plurality of candidate conversion models comprise at least a first conversion model having a first computation load and a second conversion model having a second computation load, the first computation load being smaller than the second computation load.

13. The electronic device according to claim 12 , wherein selecting the conversion model for each image patch comprises:
selecting, for each image patch, the first conversion model for the image patch responsive to determining that a difference between a second quality factor of the image patch corresponding to the second conversion model and a first quality factor of the image patch corresponding to the first conversion model is smaller than or equal to a first predetermined threshold.

14. The electronic device according to claim 12 , wherein selecting the conversion model for each image patch comprises:
selecting, for each image patch, the first conversion model for the image patch responsive to determining that a ratio of a difference between a second quality factor of the image patch corresponding to the second conversion model and a first quality factor of the image patch corresponding to the first conversion model to the first quality factor is smaller than or equal to a second predetermined threshold.

15. The electronic device according to claim 11 , wherein the first image is a first video comprising a plurality of image frames and the second image is a second video comprising a plurality of image frames, the actions further comprising:
sending the first video and the metadata to a video requester to convert the first video into the second video;
wherein the position information indicates positions of each image patch in correspondin

Assignees

• Dell Products Lp

Inventors

• Chen Qiang
• Fernandez Orellana Pedro

Classifications

• G06T7/0002

  Inspection of images, e.g. flaw detection · CPC title

• G06T2207/30168

  Image quality inspection · CPC title

• G06T2207/20021

  Dividing image into blocks, subimages or windows · CPC title

• G06T2207/10016

  Video; Image sequence · CPC title

• G06T7/11

  Region-based segmentation · CPC title