A shape matching method based on concepts tied to integral geometry is studied. The method consists of computing the distribution of random chords over the figures to be matched and comparing them by means of the Kolmogorov-Smirnov test. A simplified form of the procedure for matching figures to a circle yields a test for circularity. It is generalized to obtain a statistical test for shape matching and then modified to produce a dissimilarity measure between shapes. The circularity test is performed on 30 shapes and compared to two other circularity measures found in the literature. Experiments are also performed using the matching procedures.