Object recognition based on gabor wavelet features

The proposed method is to recognize objects from different categories of images using Gabor features. In the domain of object recognition, it is often to classify objects from images that make only limited part of the image. Hence to identify local features and certain region of images, salient point detection and patch extraction are used. Gabor wavelet features such as Gabor mean and variance using 2 scales and 2 orientations and 2 scales and 4 orientations are computed for every patch that extracted over the salient points taken from the original image. These features provide adequate resolution in both spatial and spectral domains. Thus extracted features are trained in order to get a learning model, tested and classified using SVM. Finally, the results obtained using Gabor wavelet features using 2 scales and 2 orientations and 2 scales and 4 orientations are compared and thus observed that the latter performs better than the former with less error rate. The experimental evaluation of proposed method is done using the Caltech database.