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

Methods for determining image-subject significance and training image-subject significance determining classifier and systems for same

A classifier, a remarkable technology, applied in the field of image content analysis and search, can solve the problems of high time-consuming and labor costs, and strong dependence on subjective judgments, so as to improve the ranking position, clear standards, and save time and labor costs Effect

Inactive Publication Date: 2014-05-14
ALIBABA GRP HLDG LTD
View PDF3 Cites 18 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

If a certain image is manually marked as an image with strong visual salience, it will consume a lot of time and labor costs, and it is highly dependent on human subjective judgment.

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
  • Methods for determining image-subject significance and training image-subject significance determining classifier and systems for same
  • Methods for determining image-subject significance and training image-subject significance determining classifier and systems for same
  • Methods for determining image-subject significance and training image-subject significance determining classifier and systems for same

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0025] Before judging the saliency of the image subject, it is necessary to train a classifier for judging the salience of the image subject. In this application, the classifier can be a Support Vector Machine (Support Vector Machine, SVM, which is a method of supervised learning widely used in statistical classification and regression analysis) classifier, Adaboost classifier, etc., but the present application The scope of protection is not limited thereto.

[0026] The method for judging the salient subject of an image proposed in this application does not focus on the specific position of the salient object in the image, but focuses on distinguishing the visual effect of the image containing the salient subject from the visual effect that does not contain the salient subject from the image group Rather messy image. This is a process that has nothing to do with image content and prior knowledge. By extracting visual features such as visual salience, color, edge, and texture...

Embodiment 2

[0080] The present application also provides a method for judging the saliency of an image subject, such as figure 2 shown, including the following steps:

[0081] Step S200, acquiring an image to be judged whether it is a subject saliency image.

[0082] In an example, in an online state, an image returned by keyword search or an image uploaded by a user may be obtained as an image to be judged whether it is a salient image of the subject.

[0083] Step S210, performing visual feature extraction on the image acquired in step S200 at multiple scales, the visual feature including visual salience.

[0084] Preferably, the extracted visual features may also include at least one of color features, edge features, and texture features. More preferably, the visual features include visual salience, color features, edge features, and texture features.

[0085] Preferably, the image to be judged is firstly divided into two regions, a central region and a surrounding region, and then...

Embodiment 3

[0095] In addition, the present application also provides a system for training a classifier for judging the saliency of an image subject, such as image 3 shown, including:

[0096] The sample acquisition module 300 is used to acquire A subject salient images as positive samples, and B subject non-salient images as negative samples, wherein A and B are positive integers; wherein, the function of the sample acquisition module 300 can refer to the implementation Step S100 of Example 1.

[0097] The visual feature extraction module 310 is configured to perform visual feature extraction on positive samples and negative samples at multiple scales, the visual features include visual salience. Wherein, the function of the visual feature extraction module 310 can refer to the step S110 of the first embodiment for details.

[0098] A classifier training module 320, configured to use the extracted visual features to train a classifier for judging the salience of an image subject. Wh...

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 a method and system for training an image-subject significance determining classifier, a method and system for determining image-subject significance, and a method for searching images via visual characteristics. The method for training the image-subject significance determining classifier comprises the following steps of using A images of subject significance as a positive sample, and using B images of subject non-significance as a negative sample, wherein A and B are positive integers; extracting visual characteristics, which comprise visual significance levels, from the positive and negative samples in multiple dimensions; and training the image-subject significance determining classifier by utilizing the extracted visual characteristics. According to the methods and systems of the invention, whether an image subject has significance can be rapidly and accurately determined, thereby helping image examination, screening and retrieval, etc.

Description

technical field [0001] This application relates to the field of image content analysis and search, and in particular to a method and system for training a classifier for judging the salience of image subjects, a method and system for judging the salience of image subjects, and a method for searching images using visual features method. Background technique [0002] With the development of information technology, people's needs have evolved from simple text information to image information. In view of people's increasing demand for image data query, in order to meet users' retrieval needs based on massive image information and improve the experience of using image-based Internet applications, image retrieval technology based on content analysis has become the mainstream direction of image retrieval. In image information processing tasks such as image retrieval and active vision, it is necessary to establish description and analysis of image content without any prior informat...

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): G06K9/66
Inventor 邓宇薛晖
Owner ALIBABA GRP HLDG 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

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