Processing

Please wait...

Settings

Settings

Goto Application

1. CN113870104 - Super-resolution image reconstruction

Office
China
Application Number 202010621955.4
Application Date 30.06.2020
Publication Number 113870104
Publication Date 31.12.2021
Publication Kind A
IPC
G06T 3/40
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
3Geometric image transformation in the plane of the image
40Scaling of a whole image or part thereof
G06N 3/04
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
04Architecture, e.g. interconnection topology
G06N 3/08
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
08Learning methods
CPC
G06T 3/4053
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
3Geometric image transformation in the plane of the image
40Scaling the whole image or part thereof
4053Super resolution, i.e. output image resolution higher than sensor resolution
G06T 3/4046
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
3Geometric image transformation in the plane of the image
40Scaling the whole image or part thereof
4046using neural networks
G06N 3/0454
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
04Architectures, e.g. interconnection topology
0454using a combination of multiple neural nets
G06N 3/08
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
08Learning methods
G06T 5/00
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
5Image enhancement or restoration
G06T 2207/20084
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
2207Indexing scheme for image analysis or image enhancement
20Special algorithmic details
20084Artificial neural networks [ANN]
Applicants MICROSOFT TECHNOLOGY LICENSING, LLC
微软技术许可有限责任公司
Inventors ZHENG SHUXIN
郑书新
LIU CHANG
刘畅
HE DI
贺笛
KE GUOLIN
柯国霖
BIAN JIANG
边江
LIU TIEYAN
刘铁岩
Agents 北京市金杜律师事务所 11256
Title
(EN) Super-resolution image reconstruction
(ZH) 超分辨率图像重建
Abstract
(EN) According to the implementation of the invention, a scheme for super-resolution image reconstruction is provided. According to the scheme, an input image has a first resolution is acquired. A reversible neural network is trained with the input image, wherein the reversible neural network is configured to generate an intermediate image having a second resolution and first high frequency information based on the input image, and the second resolution is lower than the first resolution. Then, an inverse network of the trained reversible neural network is used, and an output image having a third resolution is generated based on the input image and the second high frequency information subject to the predetermined distribution, wherein the third resolution is higher than the first resolution. According to the scheme, a low-resolution image obtained by an unknown downsampling method can be effectively treated, so that a high-quality high-resolution image is obtained.
(ZH) 根据本公开的实现,提出了用于超分辨率图像重建的方案。根据该方案,具有第一分辨率的输入图像被获取。利用输入图像训练来可逆神经网络,其中可逆神经网络被配置为基于输入图像生成具有第二分辨率的中间图像和第一高频信息,并且第二分辨率低于第一分辨率。随后,利用经训练的可逆神经网络的逆网络,基于输入图像和服从预定分布的第二高频信息来生成具有第三分辨率的输出图像,其中第三分辨率高于第一分辨率。该方案能够有效地处理由未知降采样方法获得的低分辨率图像,从而获得高质量的高分辨率图像。
Related patent documents