Design of rotation, illumination, and scale invariant Gabor texture descriptor for image texture analysis and retrieval

摘要

Automatic classification in medical image processing with new problems to computer-aided detection methods is attributed to the dynamics of the image acquisition conditions. Such problems request that automatic outlines can provide robust descriptions of images irrespective of camera rotation, zoom, and illumination gradients. In order to attain better retrieval, a new HIRSG feature descriptor is intended that is self-determining of illumination, sale, and rotation differences. The proposed hybrid method is developed by implementing autocorrelation Gabor filter responses on the high frequency DWT coefficients. The proposed methodology is compared with the existing methods to show the system effectiveness.