Development of biometric systems for passenger identification based on noise-resistant coding means

The paper deals with the issues of using the biometric technologies to establish identity of a passenger. The purpose of the article is to analyze the techniques of enhancing reliability of various biometric identification facilities by means of using error correction codes. The basic elements and the principle of the classical biometric system functioning are presented. On the basis of the International Civil Aviation Organization (ICAO) recommendations, the procedure features of pattern recognition are presented. The versions to adopt the biometric passenger authentication procedures are under consideration. The conclusion is drawn that with the centralized biometric databases the issues of confidentiality and information security exist. The problems are characterized by the possibility of biometric images compromise, which can potentially lead to the loss of their confidentiality and the impossibility of their further usage for personal identification. The passenger authentication procedure involving the simultaneous use of biometric parameters and contact-free SMART cards seems more reliable. SMART cards are used for distributed storage of biometric and other additional data, thus neutralizing the disadvantages of access to the centralized databases. It is shown that the subsequent step in the development of this domain is the application of biometric cryptography proposing "linking" encryption keys and passwords with the biometric parameters of the subject. Consideration is given to the principle of "fuzzy extractor" operation as one of the variants for the "biometrics-code" converter. Feasibility and necessity of upgrading the means of noise-resistant coding in the systems being studied are shown. The use of permutation decoding data algorithms capable of adequately corresponding to the particular problems of biometric identification is proposed. On the basis of the results of optical communication channels statistical modeling, the necessary and sufficient conditions for application of the permutation decoding tools for binary codes are determined. The problem to minimize memory amount for the permutation decoder cognitive map due to the permutation orbits allocation and usage of the generated loops combinations as pointers of reference plane is solved. The resulting algorithm for finding a unique orbit number and its corresponding reference plane by means of receiver formation of arbitrary parameters permutation from the set of permissible permutations is proposed.

Codes from incidence matrices of some regular graphs

We examine the [Formula: see text]-ary linear codes with respect to Lee metric from incidence matrix of the Lee graph with vertex set [Formula: see text] and two vertices being adjacent if their Lee distance is one. All the main parameters of the codes are obtained as [Formula: see text] if [Formula: see text] is odd and [Formula: see text] if [Formula: see text] is even. We examine also the [Formula: see text]-ary linear codes with respect to Hamming metric from incidence matrices of Desargues graph, Pappus graph, Folkman graph and the main parameters of the codes are [Formula: see text], respectively. Any transitive subgroup of automorphism groups of these graphs can be used for full permutation decoding using the corresponding codes. All the above codes can be used for full error correction by permutation decoding.

Codes from incidence matrices of some Bouwer graphs

We examine the [Formula: see text]-ary linear codes from incidence matrix of the Bouwer graph [Formula: see text] with vertex set [Formula: see text] and two vertices are adjacent if they can be written as [Formula: see text] and [Formula: see text] where either [Formula: see text] or [Formula: see text] differs from [Formula: see text] in exactly one position, say the [Formula: see text]th position, where [Formula: see text] All the main parameters of the codes are obtained as [Formula: see text] Also, we determine linear codes from incidence matrices of Bouwer graphs [Formula: see text] and all the main parameters of the codes are obtained as [Formula: see text] All the above codes can be used for full error correction by permutation decoding.