In this paper, we propose noise-robust features by using local directional rank coding (LDRC) for texture classification. The LDRC is based on the sign of the difference between the central pixel and its neighbors such as the local binary pattern (LBP). However, LDRC is able to collect directional information representing the rank orderof the central pixel gray level calculated in four principal orientations in  neighborhood pixel. The four ranks are combined to get  the final code. Thus, the LDRC operator produces 81 compact and discriminative features. For multi-resolution study, LDRC is calculated by altering the window size around a central pixel but the number of samples is restricted to eight neighbors by local averaging. So, in each single scale LDRC histogram, the number of bins is kept small and constant. The proposed approach has been evaluated on representative texture databases in the presence of Gaussian noise. LDRC performs better to a number of recent state-of-the-art methods.

directional rank order, texture descriptors, Gaussian noise, local binary pattern, texture classification.