The Use Of Color, Wavelet Coefficient and Texture Features

Abstract

A Query-by-Image ability often required in many multimedia or video database application. In this case the system retrieves similar images found in the database whose features are the same as the features of the query image. The key performance is in the selection of a set of features used for characterizing images in the video database. In this research a method for searching images in a video database is developed. The method accepts an image as its input and find similar images in the database. Three features are selected to characterize images in the database : color histogram, wavelet coefficients and texture. The method is tested on video database consisting eleven videos represented in RGB dan YCbCr color models. It was found that the method found all query images as long as the quality of the input image is good.