Computerized Object Recognition System Using Undecimated Wavelet Transform and Nearest Neighbour Classifier
Australian Journal of Basic and Applied Sciences, Vol. 10(2), Pages: 85-88, 2016
4 Pages Posted: 9 Jul 2017
Date Written: January 8, 2016
Object recognition is one of the main functions in human visual system. The recognition of categories of objects in images has become a central topic in computer vision. However it is a technically challenging and most important problem in computer vision due to various illuminations, size, and colour. In this paper, a robust and automated objects recognition system is presented using Undecimated Wavelet Transform (UWT) and K-Nearest Neighbour (KNN) classifier. To extract features from the given images, UWT decomposition is taken place at a predefined level of decomposition. Then, energy features are computed from all UWT sub-bands. To facilitate the recognition process, KNN classifier is employed. In order to assess the performance of the proposed approach, experiments are carried out using COIL database images. It provides satisfactory recognition accuracy of over 80% at 6th level of UWT decomposition.
Keywords: Object Recognition, Undecimated Wavelet Transform K-Nearest Neighbor Classifier and Energy
Suggested Citation: Suggested Citation