Object Recognition Based on Empirical Wavelet Transform 

Abstract

Object recognition is the method of finding an object in an image. We recognize objects without any effort easily. It is a challenging task for computer vision systems due to the size, shape, and structure of objects in an image. In this paper, an efficient object recognition system is presented based on Empirical Wavelet Transform (EWT). The energy features obtained form the EWT decomposed image is used as features for the given object. As EWT, a multiresolution analysis, the given image is decomposed at various level of decomposition and the obtained features are analyzed at each level of decomposition. The evaluation of the system is carried on Columbia Object Image Library Dataset (COIL) which consists of 100 objects captured at different orientations. The classification is done with Knearest neighbor (KNN) which gives 98.42% accuracy.

Autism Spectrum Disorder detection by an Intelligent Deep Learning Network

“…Object identification based on an empirical wavelet transform technique, in which the classification is finished using the K-nearest neighbor [10]. Switched Inductor Quasi Z Source Inverter takes use of the reactive factor in the main circuit, which boosts the voltage while simultaneously reducing the voltage ripple.…”