Unlock instant, AI-driven research and patent intelligence for your innovation.

Image annotation method based on non-Euclidean spatial difference measure

An image annotation and difference technology, applied in the field of image annotation

Active Publication Date: 2018-03-06
NANJING UNIV OF POSTS & TELECOMM
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the first assumption is not always valid, because it is generally believed that image data is distributed in the semantic feature space and exists in a complex non-Euclidean space

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 annotation method based on non-Euclidean spatial difference measure
  • Image annotation method based on non-Euclidean spatial difference measure
  • Image annotation method based on non-Euclidean spatial difference measure

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0048] The technical solutions of the present invention will be described in further detail below in conjunction with the accompanying drawings and embodiments. The following examples are implemented under the premise of the technical solution of the present invention, and detailed implementation methods and processes are provided, and the specific implementation steps are as follows:

[0049] 1 Difference measure

[0050] From the perspective of human cognition, the differences among things are the basis for constructing different categories. When constructing different categories, an affine matrix is ​​usually constructed to describe the difference between pairs of things, and these objects can be regarded as different points in a certain vector model space. There are two strategies for constructing vector models from a given dissimilarity matrix: modified Euclidean spaces and dissimilarity metric spaces. In the present invention, the difference measure between images is t...

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 provides a method for annotating images on the basis of measurement of difference among non-Euclidean spaces. The method includes firstly, computing third-order difference features of the images by the aid of a high-order statistical process so as to achieve the purpose of effectively describing correlations among the images; secondly, estimating correlation values of each tag and image contents by the aid of a maximum a posterior algorithm on the basis of a Gaussian mixture model and incremental diffusion of the difference. The method has the advantage that problems that the quantities of digital images on shared websites of social media are rapidly increased, research enthusiasm of people for image semantic retrieval is greatly promoted, but satisfactory retrieval effects cannot be realized by a large quantity of existing technologies for automatically annotating images due to gaps among low-level features of the images and high-level semantic modes can be solved.

Description

technical field [0001] The invention relates to an image tagging method in the field of computer technology, in particular to an image tagging method based on non-Euclidean space difference measurement. Background technique [0002] In recent years, the popularity of digital cameras has led to an exponential increase in the number of digital images uploaded to websites. Such a rapidly growing image database urgently needs content-based analysis techniques to effectively organize, manage and utilize these massive information resources. The key issue in building a successful content-based image retrieval system is: how to establish an accurate correspondence between the visual information at the perceptual level of the image and the annotation information at the semantic level. [0003] After searching the existing technical literature, it was found that Duygulu et al. proposed a labeling method based on machine translation, Feng et al. proposed a labeling method based on mul...

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): G06F17/30
CPCG06F16/5866
Inventor 朱松豪罗青青孙伟梁志伟
Owner NANJING UNIV OF POSTS & TELECOMM