INVESTIGATORS

Dr. Horace H S Ip, Dr. Dinggang Shen

 

BRIEF DESCRIPTION

In this project, we present a novel set of wavelet moment invariants, together with a discriminative feature selection method, for the classification of very similar objects. These invariant features are selected by a systematic method based on the discrimination measures defined for the invariant features. Using the nearest-neighbor classifier, our wavelet moment invariants achieved the highest classification rate (100%) for all four different sets tested, compared with Zernike's moment invariants and Li's moment invariants. For a test set consisting of 26 upper cased English characters, wavelet moment invariants obtained 100% classification rate when applied to 26x30 randomly-generated noisy and scaled characters, whereas Zernike's moment invariants and Li's moment invariants obtained only 98.6% and 76.7% respectively. The theoretical and experimental analyses in this project prove that the proposed method can be used to classify many types of image objects, and is particularly suitable for classifying very similar objects.

 

FUNDING AGENCY

N/A

PUBLICATIONS

1. D G Shen and Horace H S Ip, "Discriminative Wavelet shape Descriptors for Invariant recognition of 2D patterns", Pattern Recognition, Vol. 32, No.2, pp. 151-166, February, 1999.

FIGURES OR IMAGES

(13.7kb)