The methods of fractal and multifractal analysis are now widely used because they allow us to analyze and classify digital images with complex structure. We present the results of application of these methods to images obtained by sensitive crystallization method. It was developed for studying blood images and now finds increased use in analysis of medicines, determination of food and soil quality. It seems these methods are rather perspective both in comparing images from different classes and in obtaining a classifying sign for a group of images. Problems arising in images classification are also discussed.