System and method for splicing images

Patent metadata

Field	Value
Publication number	US-11416993-B2
Application number	US-202016806207-A
Country	US
Kind code	B2
Filing date	Mar 2, 2020
Priority date	Jul 14, 2016
Publication date	Aug 16, 2022
Grant date	Aug 16, 2022

Abstract

The present disclosure relates to systems and methods for image splicing. The systems and methods may acquire a first image and a second image, determine a plurality of first feature points in a first region of the first image, determine a plurality of second feature points in a second region of the second image, then match the plurality of first feature points with the plurality of second feature points to generate a plurality of point pairs. Based on the plurality of point pairs, a third region on the first image and a fourth region on the second image may be determined. Finally, a third image may be generated based on the first image and the second image, wherein the third region of the first image may overlap with the fourth region of the second image in the third image.

First claim

We claim:

1. A method for determining a plurality of feature points of an image implemented on a computing device including at least one processor and a storage, the method comprising:
decomposing the image;
generating a difference image corresponding to the decomposed image;
generating a plurality of extreme points based on the difference image and grayscale values thereof;
filtering out at least one extreme point on a sharp boundary of the image from the plurality of extreme points;
determining the plurality of feature points based on the filtered extreme points;
determining a plurality of second feature points of a second image; and
matching the plurality of feature points with the plurality of second feature points to generate a plurality of point pairs, including:
generating a plurality of initial point pairs based on the plurality of feature points and the plurality of second feature points;
acquiring coordinates of each of the plurality of initial point pairs in the image;
acquiring second coordinates of the each of the plurality of initial point pairs in the second image;
determining a slope between the coordinates and the second coordinates for the each of the plurality of initial point pairs;
generating a histogram based on the slopes and observation times of slopes; and
selecting, from the plurality of initial point pairs, the plurality of point pairs that correspond to a largest observation time of the slopes of the plurality of initial point pairs based on the histogram.

2. The method of claim 1 , wherein the decomposing the image includes:
decomposing the image based on Gaussian pyramid decomposition.

3. The method of claim 1 , wherein the generating a difference image corresponding to the decomposed image includes:
executing a convolution between the image and a Gaussian function to generate a plurality of smoothing images;
down-sampling the plurality of smoothing images to generate the decomposed image in different scales;
determining at least one difference among the plurality of down-sampled smoothing images included in the decomposed image; and
generating the difference image based on the at least one difference.

4. The method of claim 3 , the plurality of smoothing images being arranged in layers, wherein the determining at least one difference among the plurality of smoothing images includes:
determining the at least one difference between two adjacent layers of the plurality of down-sampled smoothing images.

5. The method of claim 1 , each of the plurality of extreme points includes a pixel point that has a largest absolute value of a grayscale in a region with a preset size in the image.

6. The method of claim 5 , the difference image comprising a plurality of layers, wherein the generating a plurality of extreme points based on the difference image and grayscale values thereof includes:
in each of the plurality of layers of the difference image, comparing absolute values of grayscales of pixel points with each other in a region with the preset size in the layer; and
determining a preliminary pixel point whose absolute value of a grayscale is larger than absolute values of the grayscales of other pixel points in the region as one of the plurality of extreme points.

7. The method of claim 6 , further comprising:
modifying positions of the pixel points in the region with the preset size in the layer; and
determining the plurality of extreme points based on the pixel points with the modified positions.

8. The method of claim 7 , wherein the modifying positions of the pixel points in the region with the preset size in the layer includes:
determining a displacement of a pixel point among the pixel points according to formula:

δ
=

-

∂

D
T

∂
X

⁢

(

∂
2

⁢
D

∂

X
2

)

-
1

,

where D denotes the difference image, X=(Δx, Δy, Δσ) T , Δx denotes a difference between horizontal coordinate x of the pixel point and horizontal coordinate x p of the modified pixel point, Δy denotes a difference between vertical coordinate y of the pixel point and vertical coordinate y p of the modified pixel point, and Δσdenotes a difference between σ and σ p ; and
modifying the position of the pixel point based on the displacement.

9. The method of claim 6 , wherein the generating a plurality of extreme points based on the difference image and grayscale values thereof includes:
sorting the preliminary pixel points in a descending order according to the absolute values of grayscales of the preliminary pixel points;
determining N pixel points with the highest absolute values of grayscales among the preliminary pixel points;
modifying positions of the N pixel points; and
determining the N pixel points with the modified positions as at least a part of the plurality of extreme points.

10. The method of claim 6 , further comprising:
controlling a count of the feature points and a time required to detect the feature points by controlling sizes of the different scales.

11. The method of claim 5 , the preset size of the region is associated with intensity of at least one of the plurality of feature points, a size of an area where the at least one of the plurality of feature points is located.

12. The method of claim 1 , wherein the filtering out at least one extreme point on a sharp boundary of the image from the plurality of extreme points includes:
calculating a largest principal curvature and a smallest principal curvature of each of the extreme points based on a change of surrounding pixel points; and
determining whether the extreme point is on the sharp boundary of the image according to a ratio of the largest principal curvature to the smallest principal curvature;
if the ratio of the largest principal curvature to the smallest principal curvature is larger than a preset threshold, determining that the extreme point is on the sharp boundary of the image; and
filtering out the extreme point that is determined on the sharp boundary of the image.

13. The method of claim 12 , wherein the calculating a largest principal curvature and a smallest principal curvature of each of the extreme points based on a change of surrounding pixel points includes:
determining the largest principal curvature and a smallest principal curvature f each of the extreme points according to formula:

Assignees

• Shanghai United Imaging Healthcare Co Ltd

Inventors

• Hu Yang
• Niu Jie
• Xu Liang
• Han Yecheng
• Wang Hanyu
• Bian Yueyan
• Teng Wanli
• Cui Kai
• Yang Le

Classifications

• G06T7/33 Primary

  using feature-based methods · CPC title

• G06T2207/10081

  Computed x-ray tomography [CT] · CPC title

• G06T7/0014 Primary

  using an image reference approach · CPC title

• G06T7/174

  involving the use of two or more images · CPC title

• G06T7/97

  Determining parameters from multiple pictures (depth or shape recovery from multiple images G06T7/55; stereo camera calibration G06T7/85) · CPC title