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

Image memorizing degree predication method based on sparse low-rank regression model

A technology of regression model and prediction method, applied in character and pattern recognition, complex mathematical operations, instruments, etc., to achieve the effect of reducing feature redundancy

Inactive Publication Date: 2018-04-13
TIANJIN UNIV
View PDF1 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0008] Most of the research is conducted on image attributes alone, and the research results only show that some image attributes have an impact on image memory, and few people combine various factors to automatically perform feature selection to predict image memory.

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 memorizing degree predication method based on sparse low-rank regression model
  • Image memorizing degree predication method based on sparse low-rank regression model
  • Image memorizing degree predication method based on sparse low-rank regression model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0032] An image memory prediction method based on sparse low-rank regression model, see figure 1 , the method includes the following steps:

[0033] 101: Divide the advanced semantic feature library into an advanced semantic feature library training set and an advanced semantic feature library test set, and respectively extract image memory scores corresponding to the advanced semantic feature library training set and the advanced semantic feature library test set;

[0034] 102: Taking the training set of the advanced semantic feature library and its corresponding image memory score as the learning object, using the alternate iterative method based on the augmented Lagrangian multiplier method to solve it, and training it under the algorithm framework of the sparse low-rank regression model to obtain A relationship model between high-level semantic features and image memory scores;

[0035] 103: Predict the image memory score of the advanced semantic feature library test set ...

Embodiment 2

[0046] The following is combined with specific examples, calculation formulas, figure 2 The scheme in Example 1 is further introduced, see the following description for details:

[0047] 201: Obtain a database of image memory scores;

[0048] The image memory score database used in this method is from the SUN dataset [5] 2222 images were selected as the database images, and the image memory scores corresponding to the images were obtained through the Amazon Mechanical Turk Visual Memory Game (VisualMemory Game) to form the image memory score database. A total of 665 people participated in this game. The image memory score is a continuous value from 0 to 1. The higher the value, the easier the image is to remember. Examples of database images marked with image memory scores figure 2 shown.

[0049] 202: Extract the underlying features and label information of the images in the database to form the underlying feature database and label information database respectively; ...

Embodiment 3

[0094] Combined with the specific experimental data, Figure 2 to Figure 4 , carry out feasibility verification to the scheme in embodiment 1 and 2, see the following description for details:

[0095] 1. Image memory score database:

[0096] The Image Memorability Score database used in this method contains 2222 images from the SUN dataset, which records everyday events and scenes. The memory score of the image was obtained by Amazon Mechanical Turk's Visual Memory Game (Visual Memory Game).

[0097] Two, five comparison methods;

[0098] In this experiment, this method is compared with the following five methods:

[0099] LR (Liner Regression): Use the linear prediction function to train the relationship between the underlying features and the memory score;

[0100] SVR (Support Vector Regression): support vector regression, which strings the underlying features together, and combines the RBF kernel function to learn nonlinear functions to predict image memory;

[0101]R...

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

An image memorizing degree prediction method based on a sparse low-rank regression model comprises the following steps of dividing a high-grade meaning characteristic database into a high-grade meaning characteristic database training set and a high-grade meaning characteristic database testing set, and respectively extracting image memorizing degree scores which correspond with the high-grade meaning characteristic database training set and the high-grade meaning characteristic database testing set; using the high-grade meaning characteristic database training set and the corresponding imagememorizing degree score as learning objects, solving by means of an alternative iteration method based on an augmented lagrangian method, and obtaining a relation model between the high-grade meaningcharacteristic and the image memorizing degree score through training according to an algorithm framework of the sparse low-rank regression model; and by means of the relation model which is obtainedthrough training, predicating the image memorizing degree of the high-grade meaning characteristic database testing set. The image memorizing degree prediction method has advantages of preventing image characteristic redundancy, integrating sparse regression and low-rank expression into one framework, and realizing automatic characteristic selection. Furthermore the method can be applied for predicating the image memorizing degree.

Description

technical field [0001] The invention relates to the field of image memory degree prediction, in particular to an image memory degree prediction method based on a sparse low-rank regression model. Background technique [0002] With the popularity of various mobile devices, media information is growing explosively. Among them, pictures are filled with our lives as records of daily life and mood. The memory of images is the basic ability of human cognition, and the degree of memory of different images is different. Some images live in our minds for a long time, others are quickly forgotten. Image memory is used to measure the degree to which a picture is remembered or forgotten after a period of time. It is a stable performance of an image. [1] . Like aesthetics, image emotional semantics, popularity, and other attributes of images, image memory can be estimated from the perspective of visual content analysis. [0003] low-rank representation [2] (LRR) is a means to decom...

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): G06K9/62G06F17/18
CPCG06F17/18G06F18/213G06F18/2411G06F18/214
Inventor 褚晶辉顾慧敏苏育挺井佩光
Owner TIANJIN 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

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com