Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Image super-resolution method, device and medium based on static working point

A static working point, super-resolution technology, applied in the field of deep learning, can solve problems such as limiting the expressive ability of the model, and achieve the effect of improving the generalization ability, avoiding the phenomenon of overfitting, and improving the expressive ability.

Active Publication Date: 2022-06-03
SHANDONG UNIV
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the ReLU function will set all the feature data less than 0 to zero, which will limit the expressive ability of the model

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 method, device and medium based on static working point
  • Image super-resolution method, device and medium based on static working point
  • Image super-resolution method, device and medium based on static working point

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0062]

[0065] The concept of static operating point originates from the triode amplification principle. The triode has two states, static and dynamic, the static state is

[0067] The improved VDSR model is the same as the original VDSR model in terms of network structure, except that the original ReLU function is used.

Embodiment 2

[0074] The improved VDSR model adopts the network structure of the original VDSR model, and replaces the original ReLU function with

[0076] The training data in the training set processed in the step (1) is input into the improved VDSR model set up in the step (2).

[0077] In step (3), the initial learning rate is set to 0.0001, the optimizer selects Adam, and the batchSize is set to 16.

[0080]

[0083]

Embodiment 3

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 relates to an image super-resolution method, device, and storage medium based on a VDSR model using a static operating point ReLU function, which refers to: input an image to be processed into a trained and improved VDSR model, and output it to obtain its high resolution rate image; the static operating point ReLU function in the improved VDSR model is the ReLU function of the adaptive learning static operating point; inspired by the triode amplifier circuit, the present invention introduces the concept of the static operating point into the static operating point ReLU function, and the traditional The zero point in the ReLU function is used as a static operating point, and the value of the static operating point is adaptively learned during the neural network training process. The static operating point ReLU function is applied to the VDSR model, and the data augmentation and learning rate decay strategies are adopted in the process of training the network to avoid over-fitting of the network. The invention can effectively improve the performance of the VDSR model in super-resolution tasks.

Description

A method, device and medium for image super-resolution based on static working point technical field The present invention relates to a kind of image super-resolution method based on the VDSR model of applying static working point ReLU function, The device and the storage medium belong to the technical field of deep learning. Background technique [0002] Deep learning is a collection of algorithms for modeling high-complexity data through multi-layer nonlinear transformations. With its powerful learning and expression capabilities, the network has become one of the most important research directions in the field of deep learning. It is widely used in the fields of frequency processing and so on. Each neuron node in the neural network accepts the output value of the neuron in the previous layer as the input of the neuron value and pass the input value to the next layer, the input layer neuron node will directly pass the input attribute value to the next layer (hidden...

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
Patent Type & Authority Patents(China)
IPC IPC(8): G06T3/40G06N3/04G06N3/08G06K9/62G06V10/774G06V10/82
Inventor 元辉姜东冉付丛睿姜世奇
Owner SHANDONG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products