Dual Tree Complex Wavelet Transform Based Multiclass Object Classification

Multiclass object classification is a difficult problem in computer vision application, because of highly variable nature of different objects. The primary goal of this paper is to classify object into one of the chosen classes. The proposed method uses Dual tree complex wavelet transform coefficients as a feature of object. Dual tree complex wavelet transform is having advantage of its better edge representation and approximate shift-invariant property as compared to real valued wavelet transform. We have used multiclass support vector machine classifier for classification of objects. The proposed method has been tested on dataset prepared by authors of this paper. We have tested the proposed method on multiple levels of Dual tree complex wavelet transform. Quantitative evaluation results demonstrate that the proposed method gives better performance for multiclass object classification in comparison to other state-of-the-art methods.