Recognition of Partially Occluded Objects

Research output: Journal Publications and ReviewsRGC 22 - Publication in policy or professional journal

10 Scopus Citations
View graph of relations

Author(s)

Related Research Unit(s)

Detail(s)

Original languageEnglish
Pages (from-to)228-236
Journal / PublicationIEEE Transactions on Systems, Man and Cybernetics
Volume23
Issue number1
Publication statusPublished - Jan 1993

Abstract

A computer vision system for the recognition of real world image is developed and reported. The system is capable of identifying multiple overlapped objects in a scene without stringent restrictions on their size, shape and orientation. The approach, basically, removes the dependence of an object representation to its spatial parameters with the use of difference chain code (DCC) in contour encoding. An object shape is identified by the system through the detection of selected discrete feature segments in the contour code instead of attempting to search for a complete boundary. Consequently, an object that is partially occluded can still be recognized with its remaining unmask portion. Extraction of salient features from an unknown geometry is performed using the nonlinear elastic matching technique. This algorithm is insensitive to sizing and distortions of the feature segments, hence reducing the problems caused by the error imposed during the image capturing process. Finally, a multilayer artificial neural network is employed to conclude on the identity of an unknown object based on the extracted features. The system has been applied in practice to evaluate its performance and the results are encouraging. In particular, a case study on the recognition of handtools with different surface reflectiveness are presented as an example. Possible improvements in enhancing the performance of the system are discussed. © 1993 IEEE