Title

Relational Match With Hybrid Minimum Error For Object Recognition Of Superquadrics

Keywords

Interpretation tree searching; Minimum error evaluation; Objects recognition; Relational match; Superquadric parts

Abstract

A novel method was proposed for the error in extracting relation characteristic of superquadric parts and the part-level object recognition on condition that the attribute describing of superquadrics is not provided with uniqueness. The method proposed integrated the relational match of minimum structure-error with the parts attribute constraints to implement the object recognition efficiently and accurately. Using the relative poses between the parts, the minimum error evaluation function was designed through synthesizing the parts attribute, relational attribute and structure-error. Then the recognition under the multi-object scene was carried out efficiently by searching the interpretation tree pruned.

Publication Date

4-20-2009

Publication Title

Xitong Fangzhen Xuebao / Journal of System Simulation

Volume

21

Issue

8

Number of Pages

2130-2134

Document Type

Article

Personal Identifier

scopus

Socpus ID

65649122642 (Scopus)

Source API URL

https://api.elsevier.com/content/abstract/scopus_id/65649122642

This document is currently not available here.

Share

COinS