An overview of fingerprint description models

Personal fingerprint identification is the most common method of biometric identification. In this paper we presented a review of modern fingerprint image representation models and fingerprint matching methods based on these models. Fingerprint image may be represented as a set of local features, a set of global features, a set of papillary lines and in the form of topological object which includes combination of features. The main advantages and disadvantages of common fingerprint representations were defined. Also authors proposed a classification of fingerprint represented models by the level of integrity of their description. Each level of this model describes fingerprint image using some characteristic (local features, global features, direction and density of ridges, direction field et al.) or as a solid topological object at the highest model level. Each level of the model can be used independently or with combination of other models. It allows expanding and supplementing fingerprint image description. The review of models and identification methods based on wide range of patents, scientific articles and program registration certificates for the past few years, which confirms the relevance of the problem and this research.