The two-dimensional (2-D) autocorrelation function (ACF) of an image statistically characterizes the spatial pattern within that image and presents a powerful tool for fabric analysis. It determines shape preferred orientation, degree of alignment, and distribution anisotropy of image objects. We present here a fast, user-friendly, MS-DOS based computer program, AUTO, to calculate the 2-D ACF of a digital monochrome image. AUTO displays an image on the screen and allows selection of a portion of the image for autocorrelation. Rapid calculation of ACF values is achieved by using a fast Fourier transform (FFT) routine according to the convolution theorem. The spatial distribution of ACF values is contoured for quantitative analysis of fabric anisotropy. Applications of this technique include the determination of grain or pore fabric of geological specimens, strain analysis, and the interpretation of petrophysical properties.