A set of algorithms for computing the fractal dimension from image data is introduced. The advantages and disadvantages for each algorithm are clarified. The implementation of some of these algorithms for evaluating the fractal dimension to several images is also given. Such images are considered in the present work to be generated with the Mathematica package using fractal Brownian motion. Moreover, the author generates other images by introducing a noise with different levels to a smooth gray level surface. As a result, it is concluded that the Sarkar method gives results that are more efficient and accurate than the other considered results. Our conclusion is supported by charts illustrating these results .