Image Recognition Using Combination of Multiwavelet and Radon Transforms with Neural Network


In many of the digital image processing application, observing image is modeled to be corrupted by different type of noise that result in a noisy version. Hence, image classification is an important problem that aim to find an estimate version from image have a noise that is close to the original image as possible. In the last few years, for image classification, accuracy of previous methods like Fourier transform, wavelet transform, and other methods are not so high, so they neglect some particular characters of image data. In this paper, classification method based on multiwavelet transform and radon transform that proposed, and these two transforms combine together to extract useful information from image, and then forward these features extraction for classification by using robust method of artificial neural network. The aim of this paper is that how the noisy image can be classified properly into original image via high recognition rate. A successful recognition rate of 99.3% was achieved.