Unlock instant, AI-driven research and patent intelligence for your innovation.

Image super-resolution processing method

A technology of super-resolution and processing methods, applied in the field of image processing, which can solve the problems of blurred details and inaccurate predictions.

Pending Publication Date: 2021-04-06
北京纳析光电科技有限公司
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In the field of microscopic imaging, compared with image super-resolution processing methods based on improved optical systems, although algorithm-based image super-resolution processing methods can avoid phototoxicity and photobleaching, the existing algorithm-based image super-resolution processing methods Although the high-resolution processing method has shown good results under many specific imaging conditions, there are still problems such as inaccurate prediction and fuzzy details, especially under the condition of low signal-to-noise ratio of the original data. Image super-resolution processing methods cannot achieve ideal results

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Image super-resolution processing method
  • Image super-resolution processing method
  • Image super-resolution processing method

Examples

Experimental program
Comparison scheme
Effect test

example 1

[0151] Figure 7 Shown is a comparison diagram of the low-resolution natural image according to Example 1 and the high-resolution natural image obtained after the super-resolution prediction of the first model.

[0152] Such as Figure 7 As shown, the low-resolution natural image is obtained after the super-resolution prediction of the first model (Fourier domain feature channel attention convolutional neural network model), and the high-resolution natural image is obtained. The specific steps are as follows:

[0153] Download the public dataset DIV2K, down-sample the images to generate low-resolution images corresponding to high-resolution images one-to-one, and then amplify the data, including operations such as random cropping, random angle rotation, and image symmetry, resulting in 30,000 pairs Low-resolution (dimensions 128×128×3)-high-resolution (dimensions 256×256×3) RGB image pairs (training image set, the low-resolution image corresponds to the first image, and the h...

example 2

[0160] Figure 8 Shown is a comparison diagram of the low-resolution microscopic image according to Example 2 and the high-resolution microscopic image obtained after the super-resolution prediction of the second model;

[0161] Such as Figure 8 As shown, the low-resolution microscopic image is subjected to the super-resolution prediction of the second model (Fourier domain feature channel attention generation against the convolutional neural network model) to obtain a high-resolution microscopic image. The specific steps are as follows:

[0162] Use the self-built optical microscope to take multiple sets of original images in the structured light illumination mode. In the structured light illumination super-resolution imaging mode, each area corresponds to 9 original images, and the 9 images are averaged to obtain the low-resolution wide-field illumination. image (which can be used as the input image when training the model, which is equivalent to the first image), and at t...

example 3

[0170] Figure 9 Shown is a comparison diagram of the original image illuminated by low-resolution structured light according to Example 3 and the high-resolution reconstructed image obtained after super-resolution reconstruction of the first model.

[0171] Such as Figure 9 As shown, the original image illuminated by structured light undergoes super-resolution reconstruction of the first model (Fourier domain feature channel attention convolutional neural network model) to obtain a high-resolution reconstructed image. The specific steps are as follows:

[0172] Use the self-built optical microscope to take multiple sets of original images in the structured light illumination mode, and use the traditional structured light illumination super-resolution reconstruction algorithm to perform super-resolution reconstruction on the original images taken, and obtain the image set used to form the training image set. This image set is preprocessed and augmented to generate 30,000 low...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses an image super-resolution processing method, and the method comprises the steps: obtaining a training image set which comprises a plurality of image groups, wherein each image group comprises a first image and a second image, the first image and the second image correspond to each other, and the resolution of the first image is lower than the resolution of the second image; building a first model or a second model based on a Fourier domain feature channel attention mechanism and a convolutional neural network; training a first model through the training image set, or training a second model through the training image set; and completing the super-resolution processing of a to-be-processed image through the trained first model or second model. According to the method, image features can be extracted more effectively, and super-resolution image prediction and reconstruction which are more accurate and more robust than those of an existing method are realized under different super-resolution modes and different imaging conditions.

Description

technical field [0001] The invention relates to the field of image processing, in particular to an image super-resolution processing method. Background technique [0002] In the field of microscopic imaging, fluorescence microscopic imaging is a microscopic imaging technology that can observe biological cells, tissues, and organs under living conditions, and is a powerful booster for the development of life sciences. But generally speaking, the resolution of fluorescence microscopy imaging is limited to about half of the fluorescence wavelength (250nm) by the optical diffraction limit, making traditional fluorescence microscopy imaging techniques unable to see finer biological structures than the diffraction limit. In recent years, super-resolution fluorescence microscopy imaging technology based on the improvement of optical system has developed vigorously. Stochastic optical reconstruction microscopy (STORM technology for short) can increase the resolution of optical imag...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06T3/40G06T5/10G06N3/04G06N3/08
CPCG06T3/4053G06T5/10G06N3/084G06T2207/20056G06T2207/20081G06T2207/20084G06N3/045
Inventor 李栋乔畅李迪戴琼海
Owner 北京纳析光电科技有限公司