Abstract In the paper we investigate an application of the General Shape Analysis approach to the coarse classification of teeth shapes. The main purpose of the General Shape Analysis is to find one or a few general templates with the highest similarity for an investigated object. This allows to obtain the most basic information about an object and is useful in the case of large amount of data, missing data or when classes are not evident. In this paper various shape description algorithms are applied to the problem of General Shape Analysis, namely the Two-Dimensional Fourier Descriptor, Generic Fourier Descriptor, UNL-Fourier Descriptor, Point Distance Histogram and Zernike Moments. The original description vectors are also investigated after reducing with the use of various approaches, among which are the Principal Component Analysis and Discrete Cosine Transform. We examine some methods using an original database for the General Shape Analysis. Then the selected methods are applied for the recognition of teeth shapes.