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

Image ascending frequency system and training method and image ascending frequency method thereof

A training method and image technology, which is applied in the directions of image communication, image data processing, image image conversion, etc., can solve problems such as large amount of calculation and inability to flexibly adjust the upscaling multiple, and achieve reduction in number, reduction of input signal volume, and simplification of upscaling. The effect of frequency calculation

Active Publication Date: 2015-12-02
BOE TECH GRP CO LTD
View PDF3 Cites 33 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] However, the current image upscaling system has a large amount of data calculation, and the upscaling multiple cannot be adjusted flexibly

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 ascending frequency system and training method and image ascending frequency method thereof
  • Image ascending frequency system and training method and image ascending frequency method thereof
  • Image ascending frequency system and training method and image ascending frequency method thereof

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0044] Convolutional neural network is a kind of artificial neural network, which has become a research hotspot in the field of speech analysis and image recognition. Its weight sharing network structure makes it more similar to biological neural networks, reducing the complexity of the network model and reducing the number of weights. This advantage is more obvious when the input of the network is a multi-dimensional image, so that the image can be directly used as the input of the network, avoiding the complicated feature extraction and data reconstruction process in the traditional recognition algorithm. The convolutional network is a multi-layer perceptron specially designed to recognize two-dimensional shapes. This network structure is highly invariant to translation, scaling, tilting, or other forms of deformation.

[0045] Based on the fact that the convolutional neural network does not deform the two-dimensional shape, that is, the height of the image, the present inve...

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 ascending frequency system and a training method and an image ascending frequency method thereof. Feature images of an image is obtained by a convolution neural network module, and ascending frequency processing is carried out on the image by using a synthesizer to synthesize each n*n feature images in an input signal into a feature image, whose resolution is amplified for n times. In the ascending frequency process of the synthesizer, the information of the feature images in the input signal is recorded in the generated feature image without loss, so that a high-quality image whose resolution is amplified for n times can be obtained by the synthesizer. Moreover, in the image ascending frequency system, a plurality of synthesizers can be arranged to carry out gradual ascending frequency, and each synthesizer can execute a single time of ascending frequency function, so that the ascending frequency time of the system can be flexibly adjusted according to demand. Further, when the synthesizer amplifies the resolution of the feature image of n times, the number of the feature images output by the synthesizer is reduced, thereby reducing the signal input flow of the next synthesizer or the first convolution neural network module and simplifying the calculated amount thereof.

Description

technical field [0001] The present invention relates to the technical field of image signal processing, in particular to an image upscaling system, its training method and an image upscaling method. Background technique [0002] At present, in the process of image signal processing, the image resolution is generally improved by using standard upscaling (improve image resolution) methods such as bicubic and linear. Such as figure 1 As shown, a 2x upscaling method is shown, using four different filters F1, filter F2, filter F3 and filter F4 for each pixel of the input image (plus adjacent pixels), each filter Produces a quarter of the pixels of the output image, this process can be seen as applying 4 filters (convolution) to the input image and then interleaving or multiplexing to create a single output image with double the width and height. [0003] However, the current image upscaling system has a large amount of data calculation, and the upscaling multiple cannot be adju...

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 Applications(China)
IPC IPC(8): H04N5/14
CPCG06T3/4053G06N3/045G06F18/251G06T3/40G06N3/04G06N3/08
Inventor 那彦波张丽杰何建民
Owner BOE TECH GRP CO LTD
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