Method and apparatus with blur estimation

Patent metadata

Field	Value
Publication number	US-11636577-B2
Application number	US-202016867749-A
Country	US
Kind code	B2
Filing date	May 6, 2020
Priority date	Jun 27, 2019
Publication date	Apr 25, 2023
Grant date	Apr 25, 2023

Abstract

A processor-implemented method with blur estimation includes: acquiring size information of an input image; resizing the input image to generate a target image of a preset size; estimating a blur of the target image; and estimating a blur of the input image based on the size information of the input image.

First claim

What is claimed is:

1. A processor-implemented method with blur estimation, comprising:
acquiring size information of an input image;
detecting a region of interest (ROI) in the input image;
resizing the ROI to generate a target image of a preset size;
estimating a blur of the target image; and
estimating a blur of the input image based on the estimated blur of the target image and the size information of the input image.

2. The method of claim 1 , wherein the acquiring of the size information of the input image comprises acquiring size information of the ROI in the input image.

3. The method of claim 1 , wherein the estimating of the blur of the input image comprises correcting the blur of the input image based on a ratio between the size information of the input image and size information of the target image.

4. The method of claim 3 , wherein each of the size information of the input image and the size information of the target image comprises an area, a height, and a width.

5. The method of claim 1 , further comprising:
performing a liveness test on the input image based on a result of a comparison between the estimated blur of the input image and a threshold.

6. The method of claim 5 , wherein the threshold adaptively varies based on the size information of the input image.

7. The method of claim 5 ,
wherein, in response to the estimated blur of the input image being less than the threshold, a result of the liveness test is that the input image represents a live image, and
wherein, in response to the estimated blur of the input image being greater than or equal to the threshold, the result of the liveness test is that the input image does not represent the live image.

8. The method of claim 1 , wherein the ROI includes a facial landmark for each of plural discrete portions of a face.

9. The method of claim 1 , wherein the estimating of the blur of the input image comprises estimating the blur of the input image using a blur estimation machine learning model.

10. A non-transitory computer-readable recording medium storing instructions that, when executed by a processor, cause the processor to perform the method of claim 1 .

11. The method of claim 9 , wherein the blur estimation machine learning model is a blur estimation neural network.

12. An apparatus with blur estimation, comprising:
a memory storing computer-readable instructions; and
one or more processors,
wherein, in response to the instructions being executed by the one or more processors, the one or more processors are configured to:
acquire size information of an input image;
detect a region of interest (ROI) in the input image;
resize the ROI to generate a target image of a preset size;
estimate a blur of the target image; and
estimate a blur of the input image based on the estimated blur of the target image and the size information of the input image.

13. The apparatus of claim 12 , wherein the one or more processors are further configured to acquire the size information of the input image by acquiring size information of the ROI in the input image.

14. The apparatus of claim 12 , wherein the processor is further configured to estimate the blur of the input image by correcting the blur of the input image based on a ratio between the size information of the input image and size information of the target image.

15. The apparatus of claim 14 , wherein each of the size information of the input image and the size information of the target image comprises an area, a height, and a width.

16. The apparatus of claim 14 , wherein the estimating of the blur of the input image is further based on the preset size.

17. The apparatus of claim 12 , wherein the one or more processors are further configured to perform a liveness test on the input image based on a result of a comparison between the estimated blur of the input image and a threshold.

18. The apparatus of claim 17 , wherein the threshold adaptively varies based on the size information of the input image.

19. The apparatus of claim 17 ,
wherein, in response to the estimated blur of the input image being less than the threshold, a result of the liveness test is that the input image represents a live image, and
wherein, in response to the estimated blur of the input image being greater than or equal to the threshold, the result of the liveness test is that the input image does not represent the live image.

20. The apparatus of claim 12 , wherein the ROI includes a facial landmark for each of plural discrete portions of a face.

21. The apparatus of claim 12 , wherein the one or more processors are further configured to estimate the blur of the input image using a blur estimation machine learning model.

22. An apparatus with user verification, comprising:
a camera configured to capture an input image; and
one or more processors configured to:
detect a region of interest (ROI) in the input image;
resize the ROI to generate a target image of a preset size different from a size of the input image;
estimate a blur of the target image;
estimate a blur of the input image based on the size of the input image, the preset size, and the estimated blur of the target image;
perform a liveness test on the input image based on the estimated blur of the input image; and
perform the user verification based on a result of the liveness test.

23. The apparatus of claim 22 , wherein the preset size is smaller than the size of the input image, and each of the preset size and the size of the input image is a resolution.

24. The apparatus of claim 22 , wherein the estimating of the blur of the input image based on the size of the input image, the preset size, and the estimated blur of the target image comprises:
estimating the blur of the input image by determining a ratio between the size of the input image and the size of the target image; and
applying the determined ratio to the estimated blur of the target image.

25. The apparatus of claim 22 , wherein the performing of the liveness test on the input image based on the estimated blur of the input image comprises comparing the estimated blur of the input image to a threshold.

26. The apparatus of claim 25 , wherein, in the performing of the user verification based on the result of the liveness test,
in response to the estimated blur of the input image being greater than or equal to the threshold, a result of the user verification is that the user fails verification, and
in response to the estimated blur of the input image being less than the threshold, the result of the user verification is that the user is verified.

27. The apparatus of claim 25 , wherein the threshold adaptively varies based on size information of the input image.

28. The apparatus of claim 22 , wherein the size of the input image is an acquired size information of the ROI in the input image.

Assignees

• Samsung Electronics Co Ltd

Inventors

• Park Seong-Jin
• Baek Jiwon
• Han Seungju
• Ko Minsu
• Lee Solae
• Lee Hana

Classifications

• G06T7/11

  Region-based segmentation · CPC title

• G06T7/62

  of area, perimeter, diameter or volume · CPC title

• G06V10/82

  using neural networks · CPC title

• G06T7/337

  involving reference images or patches · CPC title

• G06V10/25

  Determination of region of interest [ROI] or a volume of interest [VOI] · CPC title