Saliency compression method and system, storage medium, computer equipment and application

A saliency, image compression technology, applied in computer components, computing, neural learning methods, etc., can solve problems such as poor saliency compression algorithms

Pending Publication Date: 2021-07-30
XIDIAN UNIV
View PDF2 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] Aiming at the problems existing in the prior art, the present invention provides a saliency compression method, system, storage medium, computer equipment and application, aiming to solve the problem that the existing saliency compression algorithm is not effective

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
  • Saliency compression method and system, storage medium, computer equipment and application
  • Saliency compression method and system, storage medium, computer equipment and application
  • Saliency compression method and system, storage medium, computer equipment and application

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 2

[0132] A method of significant compression provided by an embodiment of the present invention, including:

[0133] Step 1, the data set The present invention adopts the DUTS-TR dataset as the training data set, which contains 10,533 pictures. This dataset is a significant detection data set in the data set. For test data sets, the present invention uses six common reference data sets, including:

[0134] Dut-OMRON includes 5168 images, most of which contain a small amount of complicated foreground objects. DUTS-TE contains 5019 images. The HKU-IS contains 4447 images in which multiple foreground images are included. ECSSD contains 1000 complex images, and many images contain large foreground objects. Pascal-s contains 850 outlook complex images with messy background. The SOD contains only 300 images. However, the front border of the picture is blurred, overlapping, and its contrast is low, and the segmentation effect is poor.

[0135] Step 2, training settings. In addition to the ...

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 belongs to the technical field of image compression, and discloses a saliency compression method and system, a storage medium, computer equipment and application, and the method comprises the following steps: for a saliency detection module, employing a U2Net network to achieve the saliency detection module; comparing a residual block fusion local feature with a multi-scale feature, and proposing an overall network architecture; according to the U-Net and the non-local residual error, proposing ResU to extract multi-scale features in the stage; inputting a convolutional layer, and converting the original feature map into an intermediate map F1 (x) through a common convolutional layer; learning and coding multi-scale context information through U-block by taking the intermediate feature map F1 (x) as input; fusing local features and multi-scale features through F1 (x) and U (F1 (x)); for a compression module, proposing an attention channel method, so that a discrete Gaussian model is improved, and finally decoder enhancement is introduced to realize construction of an image compression architecture. According to the invention, the compression quality of the image can be improved at a low bit rate.

Description

Technical field [0001] The present invention belongs to the field of image compression, and more particularly to a method, system, storage medium, computer device, and application of a significant compression. Background technique [0002] Currently, although the amount of data in the picture is large, not all content contains user interest. People often pay more attention to the goal in typical application scenarios. The identification problem of the significant target is particularly important under the conditions of bandwidth limit. The global compression method will make some important details of the significant region, such as license plate numbers, indications, etc. The compression ratio of the significant region can reduce the compression ratio of the significant region, improve the compression magnification of the irrelevant information such as the background, and under the same picture compression, the information that is interested in users can be reserved as much as po...

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): G06T9/00G06K9/62G06N3/04G06N3/08
CPCG06T9/002G06N3/08G06N3/045G06F18/253
Inventor 赵楠孙向前
Owner XIDIAN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap