Gabor filters for rotation invariant texture classification
Robert P. Porter, Nishan C. Canagarajah · 1997
A Gabor filter based feature extraction scheme for texture classification is proposed. By using a novel set of circularly symmetric filters, rotation invariance is achieved. The scheme offers a high classification performance on textures at any orientation using both fewer features and a smaller area of analysis than most existing schemes. The performance of the scheme on noisy images is also investigated, demonstrating a high robustness to noise.