Informatica

Filtering of feature matches is heuristic method aimed to reduce the number of feasible matches and is widely employed in different image registration algorithms based on local features. In this paper we propose to interpret the filtering process as an optimal classification of the matches into the correct or incorrect match classes. The statistics, according to which the filtering is performed, uses differences of the geometrical invariants obtained from ordered sets of local features (composite features) of proper cardinality. Further, we examine some computationally efficient implementation schemes of the classification. Under the assumption of Gaussian measurement error, the conditional distribution densities of invariants can be approximated by well-known linearization approach. Experimental evidences obtained from fingerprint identification, which confirm viability of the proposed approach, are presented.

Fingerprint ridge frequency is a global feature, which is most prominently different in fingerprints of men and woman, and it also changes within the maturing period of a person. This paper proposes the method of fingerprint pre-classification, based on the ridge frequency replacement by the density of edge points of the ridge boundary. This method is to be used after applying the common steps in most fingerprint matching algorithms, namely the fingerprint image filtering, binarization and marking of good/bad image areas. The experimental performance evaluation of fingerprint pre-classification is presented. We have found that fingerprint pre-classification using the fingerprint ridge edges density is possible, and it enables to preliminary reject part of the fingerprints without heavy loss of the recognition quality. The paper presents the evaluation of two sources of fingerprint ridge edges density variability: a) different finger pressure during the fingerprint scanning, b) different distance between the geometrical center of the fingerprint and position of the fingerprint fragment.

The paper presents a fingerprint registration approach based on the decomposition of registration process into elementary stages. In each stage a single transformation parameter is eliminated. The algorithm uses composite features, i.e., lines connecting two minutiae instead of fingerprint minutiae. These features have rotation and translation-invariant attributes allowing feature filtering with significantly enhanced signal-to-noise ratio in feature consensus scheme. Experimental results of goal-directed performance evaluation with live-captured fingerprint image database are presented.

The principles of a neural network environmental model are proposed. The principles are universal and can use different neural network architectures. Such a model is self-organizing, it can operate in both regimes with and without a teacher. It codes information about objects, their features, the actions operating in an environment, analyzes concrete situations. There are functions for making an action plan, for action control. The goal of the model is given from an external site. The model has more than sixteen active regimes. The neural network environmental model is fulfilled in software and hardware tools.