Comparative analysis using fast discrete Curvelet transform via wrapping and discrete Contourlet transform for feature extraction and recognition

In this paper, comparative analysis for feature extraction and recognition based on fast discrete Curvelet transform via wrapping and discrete Contourlet transform using Neural Network and Euclidean distance classifier is proposed. The pre processing is applied on the each image of dataset. Each image from the Training Dataset is decomposed using the fast discrete Curvelet transform and discrete Contourlet transform. The Curvelet coefficients as well as Contourlet coefficients of low frequency & high frequency in different orientation and scales are obtained. The frequency coefficients are used as a feature vector for further process. The PCA (Principal component analysis) is used to reduce the dimensionality of the feature vector. Finally the reduced feature vector is used to train the Classifier. The test databases are projected on Curvelet-PCA and Contourlet-PCA subspace to retrieve reduced coefficients. These coefficients are used to match the feature vector coefficients of training dataset using Neural Network Classifier. The results are compared with the results of Euclidean distance classifier for both the methods.