Classification of Image by Combining Wavelet Transform and Neural Network
Dharmendra Patidar1, Nitin Jain2
1Dharmendra Patidar, Electronics and Telecommunication, R.G.P.V, Mandsaur, Mandsaur, India.
2N. Nitin Jain, Electronics and Telecommunication, R.G.P.V, Mandsaur,  Mandsaur, India.
Manuscript received on September 29, 2013. | Revised Manuscript received on October 12, 2013. | Manuscript published on October 30, 2013. | PP: 404-408  | Volume-3, Issue-1, October 2013. | Retrieval Number:  A2306103113/2013©BEIESP

Open Access | Ethics and Policies | Cite
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (

Abstract: Image classification plays an important role in many tasks, which is still a challenging problem in organizing a large image database. However, an effective method for such an objective is still under investigation. In this paper, we propose a supervised method for image classification based on combination of wavelet transform and Neural Network (NN). Neural network has been increasingly used in image classification in the last few decades. The proposed scheme for successful classification is combination of a wavelet domain feature extractor and back propagation neural networks (BPNN) classifier. For achieving a suitable way for classification of image here we first use wavelet transform. In present day wavelet transform is most popular and widely used method for image classification. Wavelet transform is a well-known tool for signal/image analysis. It provides a time–frequency representation of the data as well. Wavelet transform first takes image from given data base, analysis this image and decompose main image into sub image and gives information about texture and shape from given image. In this proposed method of image classification first we divide all given image into six parts. For obtaining the necessary and required information from each part of the given divided image we use first order color movements and daubechies4 (db4) types of wavelet transform. This proposed method for classification of image is fully based on back propagation. Information about the color movement is used as a first input for NN. Second input is a deubechies4 transform of wavelet is used for NN. Final step of classification is based on back propagation neural network (BPNN) with one hidden layer. Back propagation, an abbreviation for “backward propagation of errors”, is a common method of training artificial neural networks. backpropagation is based on weight of input and output neurons. In neuroscience and computer science, synaptic Weight refers to the strength or amplitude of a connection between two nodes, corresponding in biology to the amount of influence the firing of one neuron has on another. The term is typically used in artificial and biological neural network research this new approach of classification of image is based on the texture, information of color and shape.170 aircraft color image were used for training and 200 for testing. Resulting data consist of 98% and 90% efficiency for training and testing respectively.
Keywords: Back Propagation, Color Moment, Neural Network, Wavelet Transform.