Texture classification is widely employed in many computer vision and pattern recognition applications. Texture classification is performed in two phases namely feature extraction and classification. Several feature extraction methods and feature descriptors have been proposed and local binary pattern (LBP) has attained much attraction due to their simplicity and ease of computation. Several variants of LBP have been proposed in literature. This paper presents a performance evaluation of LBP based feature descriptors namely LBP, uniform LBP (ULBP), LBP variance (LBPV), LBP Fourier histogram, rotated LBP (RLBP) and dominant rotation invariant LBP (DRLBP). For performance evaluation, nearest neighbor classifier is employed. The benchmark OUTEX texture database is used for performance evaluation in terms of classification accuracy and runtime.