Performance analysis of K Nearest Neighbors image classifier with different wavelet features

Abstract

Classification of mutually class image plays most important role in different engineering and computer vision application. Some important fields where these types of classification technique are widely used include image processing in medical, robotics based on classification, pattern recognition. Successful image classification is very challenging task especially when image database is very large. To solve this challenging task scientists and researchers have made a lot of efforts and continuously work to implement a successful classification algorithm [1, 2, 3, 4, 5, and 6]. To successfully classified an image from large database in a short interval of time. Our proposed K Nearest Neighbors (KNN) classification method is based on haar, daubechies4 (db4), and discrete Mayer (demy) wavelet features. In this proposed method classification accuracy of different wavelet features are comparing by KNN classifier in terms of classification efficiency. Proposed work is completely experimented on Matlab 2011b software and this work present a new application and contribution towards image classification.