Deformable part model object detection method based on color description

A technology for deforming parts and color description, applied to computer parts, character and pattern recognition, instruments, etc., can solve problems such as ignoring the description of image color features

Active Publication Date: 2014-11-05
BEIJING UNIV OF TECH
View PDF2 Cites 31 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, like most of the current state-of-the-art object detection methods, the deformable part model is only based on the grayscale feature for detection, ignoring the description of the color feature of the image

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
  • Deformable part model object detection method based on color description
  • Deformable part model object detection method based on color description
  • Deformable part model object detection method based on color description

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0052] In order to enable those skilled in the art to better understand and use the present invention, the technical solutions of the present invention will be further described below in conjunction with the accompanying drawings and specific implementation examples.

[0053] 1. The training database contains a labeled positive sample set P and an unlabeled negative sample set N. Sort the labeled boxes in the positive sample set P according to their aspect ratio and classify them as P 1 ,...,P m groups, and the bounding boxes within each group have similar aspect ratios. Here, the aspect ratio is taken as an indicator of appearance variation for distinguishing objects of the same class. A total of m different root filters are trained. Calculate the root filter CN-HOG feature vector, the main steps are as follows: image grayscale, color space normalization, pixel gradient calculation, gradient histogram to specify weight projection, contrast normalization and truncation, cal...

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 deformable part model object detection method based on color description, and belongs to the technical field of image target detection. The invention provides an intelligent object detection method which combines shape and color features. A deformable part model is used as a bottom frame, and a linguistics-based Color Name color descriptor is added into the original feature space of a histogram of oriented gradient when a template is trained to obtain a shape template of a specific object type and a color template of the specific object type; and finally, in a detection stage, an object is detected by a sliding window method characterized by the matching of dual templates, i.e. the shape template of the histogram of oriented gradient and the Color Name color template. The defect of wrong detection since the object is described by a single feature in a traditional method is overcome.

Description

technical field [0001] The invention belongs to the technical field of image target detection, uses a deformable part model (DPM, Deformable Part Model) as the underlying framework, and adds language-based In the detection stage, the gradient direction histogram shape template and the color template double-template matching sliding window method are used to detect the target. The present invention not only makes use of the part model and the multi-view mixed model in the traditional deformable part model to make the detection have the flexibility similar to that of human beings to recognize objects, but also because of the introduction of color description, the method of the present invention has the diversity of features similar to human observation of objects, These factors narrow the semantic gap that exists in object detection. Background technique [0002] In recent years, the research on service robot technology has made great progress, and various types of service ro...

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): G06K9/62G06K9/66G06K9/46
Inventor 杨金福张济昭高晶钰张珊珊李明爱张强陈浩
Owner BEIJING UNIV OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap