Image enlargement method and apparatus

Patent metadata

Field	Value
Publication number	US-9652825-B2
Application number	US-201414503948-A
Country	US
Kind code	B2
Filing date	Oct 1, 2014
Priority date	Dec 31, 2013
Publication date	May 16, 2017
Grant date	May 16, 2017

Abstract

The present application discloses an image enlargement method and apparatus. The method includes: inputting a low-resolution image, where the low-resolution image includes at least one matching areas; searching one of the matching areas for an optimal contour template, where the optimal contour template is a contour template with a minimum total variation value in the matching area and the contour template set; finding a second pixel point according to a zoom ratio and a first pixel point in the matching area; performing local interpolation for the second pixel point according to the optimal contour template, so as to obtain an intermediate pixel value; repeating the foregoing steps until processing for all the matching areas is complete, and obtaining an initial high-resolution image according to all of the intermediate pixel values; and performing global interpolation for the initial high-resolution image to obtain a high-resolution image.

First claim

What is claimed is:

1. An image enlargement method, the method comprising:
inputting a low-resolution image, wherein the low-resolution image comprises matching areas and the matching areas are used to match all contour templates in a contour template set;
searching one matching area for an optimal contour template;
finding a second pixel point according to a zoom ratio and a first pixel point in the matching area, wherein the second pixel point is corresponding to the first pixel point;
performing local interpolation for the second pixel point according to the optimal contour template, so as to obtain an intermediate pixel value;
repeating the steps of searching one matching area, finding a second pixel point, and performing local interpolation until processing for all the matching areas is complete, and obtaining an initial high-resolution image according to all of the intermediate pixel values;
performing global interpolation for the initial high-resolution image to obtain a high-resolution image; and
wherein performing local interpolation for the second pixel point according to the optimal contour template comprises:
calculating an oriented function of an interpolation of the second pixel point according to an optimal contour template of a matching area of the first pixel point corresponding to the second pixel point, and
performing the local interpolation for the second pixel point according to the oriented function of the interpolation of the second pixel point.

2. The method according to claim 1 , wherein the step of calculating an oriented function of an interpolation of the second pixel point according to an optimal contour template of a matching area of the first pixel point corresponding to the second pixel point comprises:
calculating a direction angle function θ S n and an anisotropy function u S of the second pixel point according to the optimal contour template of the matching area of the first pixel point corresponding to the second pixel point, wherein the oriented function ρ s*(k) n is indicated by the direction angle function θ S n and the anisotropy function u S ,

θ
S
n

=

∠
⁢

∫

∇

φ
S
⊥

⁡

(

k
-
n

)

⁢

ⅆ
x

u
S

=

min

n
∈
N

⁢



∫

∇

φ
S

⁡

(

k
-
n

)

⁢

ⅆ
x



wherein φ S is the optimal contour template, ∇φ S ⊥ indicates that φ S is projected to a first axis and a second axis that are perpendicular to each other and then gradients are separately calculated in the directions of the first axis and the second axis that are perpendicular to each other, k is the first pixel point, N indicates a pixel point set in a neighbor area of the pixel point k, and n is a pixel point in the pixel point set.

3. The method according to claim 1 , wherein the step of performing the local interpolation for the second pixel point according to the oriented function of the interpolation of the second pixel point comprises:
performing the local interpolation for the second pixel point according to

u
k

⁡

(
x
)

=

v
k

+

∑

n
∈
N

⁢

c
n

⁢

ρ

s
*

⁡

(
k
)

n

⁡

(

x
-
n

)

,

wherein x is the second pixel point, k is the first pixel point, u k (x) is the intermediate pixel value of the second pixel point, v k is a pixel value of the first pixel point, N is a pixel point set in a neighbor area of x, n is a pixel point in the pixel point set, c n is an interpolation coefficient, and ρ s*(k) n is an oriented function of an interpolation pixel point when an optimal contour template S*( ) is used.

4. An image enlargement method, comprising:
inputting a low-resolution image, wherein the low-resolution image comprises matching areas and the matching areas are used to match all contour templates in a contour template set;
searching one matching area for an optimal contour template;
finding a second pixel point according to a zoom ratio and a first pixel point in

Assignees

• Huawei Tech Co Ltd

Inventors

• Yang Xiaowei
• Li Xin
• Xiao Jinsheng

Classifications

• G06T3/4076

  using the original low-resolution images to iteratively correct the high-resolution images · CPC title

• G06T3/4069

  by subpixel displacements · CPC title

• G06T3/403 Primary

  Edge-driven scaling; Edge-based scaling · CPC title

• G06T3/4007 Primary

  based on interpolation, e.g. bilinear interpolation (image demosaicing G06T3/4015; edge-driven or edge-based scaling G06T3/403) · CPC title

• G06T3/4061

  by injecting details from different spectral ranges · CPC title