Noise robust Laws' filters based on fuzzy filters for texture classification

y Laws' mask method has achieved wide acceptance in texture analysis, however it is not robust to noise. Fuzzy filters are well known for denoising applications. This work proposes a noise-robust Laws' mask descriptor by integrating the exiting fuzzy filters with the traditional Laws' mask for the improvement of texture classification of noisy texture images. Images are corrupted by adding Gaussian noise of different values. These noisy images are transformed into fuzzy images through fuzzy filters of different windows. Then the texture features are extracted using Laws' mask descriptor. To investigate the proposed techniques two texture databases i.e. Brodatz and STex are used. The proposals are assessed by comparing the performance of the traditional Laws' mask descriptor alone and after combined with the fuzzy filters on noisy images. The k-Nearest Neighbor (k-NN) classifier is utilized in the classification task. Results indicate that the proposed approach delivers higher classification accuracy than the traditional Laws' mask method. Hence, validate that the suggested methods significantly improve the noised texture classification. (C) 2019 Production and hosting by Elsevier B.V. on behalf of Faculty of Computers and Information, Cairo University.