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

Detection device and method for multi-class targets

A detection device and detection method technology, which can be applied to instruments, character and pattern recognition, computer parts and other directions, and can solve problems such as difficulty in classifier training.

Inactive Publication Date: 2010-10-06
SONY GRP CORP
View PDF7 Cites 36 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the algorithm proposed in the same literature [2] forces feature sharing among various categories, which makes it difficult for a classifier to share features with other categories when a certain category in the multi-category cannot share features. Further training brings difficulties

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
  • Detection device and method for multi-class targets
  • Detection device and method for multi-class targets
  • Detection device and method for multi-class targets

Examples

Experimental program
Comparison scheme
Effect test

no. 2 example

[0057] According to the second embodiment of the present invention, the detection device for detecting multi-category target data is designed as a cascade structure (Cascade) classifier in which a plurality of classifiers are connected in series. To this end, the classifiers at all levels of the cascade classifier (SC k ) is artificially designed as a predetermined multi-layer structure (the first multi-layer structure in the present invention). Set the most detailed categories (for example, r categories, r is a natural number greater than 1) at the bottom, and then merge these categories into a few larger categories at a higher level according to a predetermined similarity standard, Then it is merged step by step up to the highest level, such as a large category.

[0058] Figure 5a and 5b Shows the use of a class tree structure CT to represent sample class changes during training. Figure 5a , a total of 7 types of object samples participate in the training, these 7 type...

no. 3 example

[0080] In the third embodiment, cars, buses and trucks are used as objects to be detected, and a more detailed classification (training) method for cascaded classifiers is described.

[0081] First, prepare three types of positive sample sets (car images) P(C i )(i=1, 2, 3) correspond to cars, buses and trucks respectively, and combine the three types of positive samples into one type of positive sample set P(C 0 ), the structure of the sample category tree is as follows Figure 5b shown; training from P(C i )(i=0), when it is necessary to split the positive sample category, P(C i )(i=0) is split into P(C i )(i=1, 2, 3); and set all kinds of expected training targets: detection rate D i and the total false detection rate F i ;

[0082] Next, prepare the feature pool, and apply the Haar-like feature prototype to a 32×32 (pixel) image to obtain hundreds of thousands of specific features.

[0083] Then train classifiers SC at all levels step by step 1 to SC n . Such as ...

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 relates to detection device and method for multi-class targets. The detection device comprises an input unit, a joint classifier and a discrimination unit, wherein the input unit is configured to input data to be detected; and the joint classifier comprises multiple strong classifiers which can treat multiple classes of target data, wherein each strong classifier is formed by summarizing a group of weak classifiers, and each weak classifier carries out weak classifying on the data to be detected by using one feature; the discrimination unit is configured to discriminate which class of target data the data to be detected belongs to based on the classifying results of the multiple strong classifiers, the joint classifier contains a shared feature list, and each feature in the list is respectively shared by one or more weak classifiers belonging to different strong classifiers; and the weak classifiers which have the same feature but belong to different strong classifiers have mutually different parameter values.

Description

technical field [0001] The invention relates to target detection technology. In particular, it relates to a detection device and a detection method for detecting multiple types of target data. Background technique [0002] It is becoming more and more important to use machine learning methods to detect target data on images or other data to be detected. Especially object detection in images has become one of the important branches. [0003] Affected by multiple factors such as illumination, viewing angle, and posture, the same type of object may have a huge difference in the state in the image, which brings great difficulties to the object detection technology in the image. The same type of objects may therefore be divided into multiple sub-categories for processing, but how to effectively utilize the commonality among multiple sub-categories while accurately distinguishing the differences is still a topic that needs further research. [0004] For multi-class image object...

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
CPCG06K9/6257G06F18/2148
Inventor 梅树起吴伟国
Owner SONY GRP CORP
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