Feature Extraction and Matching using Gabor wavelets and Hessian detector

The research work involves the use of two dimensional Gabor wavelets in image processing. The key idea is to utilize a Gabor wavelet as a feature extractor. By using Gabor wavelets the image is reconstructed and required features are only represented. The corners of the input image are obtained and are matched to that of the rotated image of the input. A powerful feature extractor Hessian interest point detector is compared to detectors like Derivative of Gaussian function. Most of research applications like navigation, tracking, optical flow demand perfect feature extraction and matching though the image is rotated. In the present work while capturing the image the camera is calibrated at an angle of 25 °. It is observed that the features can be perfectly extracted even after the rotation of the images. A real-time system can be developed by implementing the proposed approach on a SoC and hence the area, power and delay can be optimized