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

Trademark image identification method

An image recognition and trademark technology, applied in the field of visual target detection and recognition, can solve the problems of no filtering, no false matching elimination method, etc., and achieve the effects of fast recognition speed, reduced time complexity, and improved recognition accuracy

Active Publication Date: 2016-06-22
PEKING UNIV
View PDF5 Cites 7 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] To sum up, there are two deficiencies in the existing trademark recognition technology: one is that there is no false match elimination method that is robust to image affine transformation and missing key points; A method that can effectively filter those key points that are not related to the trademark

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
  • Trademark image identification method
  • Trademark image identification method
  • Trademark image identification method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0027] The present invention will be described in further detail below in conjunction with the accompanying drawings and specific examples.

[0028] The present invention is a trademark image recognition method, and its technical process is as follows: figure 1 and figure 2 As shown, it specifically includes the following steps:

[0029] (1) Prepare positive samples and extract visual features

[0030] Prepare a certain number of training samples for each trademark to be recognized, and extract a local feature from these training samples, such as SIFT feature or SURF feature, and then cluster and quantify the extracted features to obtain the visual word representation of key points , as shown in Formula 1:

[0031] Formula 1: k={P(k), S(k), I(k)}

[0032] Among them, k represents the kth key point, P(k) represents the position, S(k) represents the scale, and I(k) represents the nearest cluster center.

[0033] (2) Feature selection based on mutual information

[0034] F...

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

A trademark image identification method provided by the present invention comprises the following steps of preparing a training sample of a to-be-identified trademark, extracting the local characteristics of the training sample, and carrying out the clustering and quantification; utilizing the mutual information to select the characteristics; extracting the local characteristics of a tested image same as the local characteristics of the training sample, and utilizing a clustering center obtained previously to quantize the local characteristics; utilizing the characteristics retained by the characteristic selection to filter the key points of a tested picture, and carrying out the key point matching on the tested picture and a positive sample; utilizing a spatial topological relation to restrain the matching, removing the key point pairs of the wrong matching, and gathering the number of different visual words in the point matching pairs as the similarity of the tested image and the positive sample; and according to the similarity, determining whether the tested image contains the trademark. The trademark image identification method of the present invention considers from the two aspects of irrelevant characteristic point filtering and the wrong matching elimination, the irrelevant characteristic point filtering and the wrong matching elimination are the complementary relation and can be promote mutually, thereby improving the trademark identification accuracy better.

Description

technical field [0001] The invention belongs to the technical field of visual target detection and recognition, and in particular relates to a trademark image recognition method. Background technique [0002] In recent years, with the rapid development and popularization of Internet technology, especially the continuous promotion and application of social networking sites, images on the Internet have shown a rapid growth trend, which makes online media the most potential advertising and business platform. Brand tracking is a service that has emerged in recent years. It evaluates brand growth by analyzing the frequency of brand exposure in the media and user evaluations. There are two traditional analysis methods. One is to rely on manual analysis and statistics. Due to the huge amount of online media, this method requires a lot of manpower and time; the second method is to retrieve relevant information through the keyword search function of the website. However, the keyword...

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): G06K9/62
CPCG06F18/23
Inventor 唐攀攀彭宇新
Owner PEKING 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