Speciality:
05.13.18 (Mathematical modelling, calculating methods, and the program systems)
Birth date:
25.04.1952
Website: https://math.nsc.ru/~kelmanov/index.htm Keywords: pattern recognition,
operations research,
signal processing.
UDC: 519.6, 519.2, 519.1, 519.7, 621.391, 519.71
Subject:
Scientific concerns:
1) Mathematical methods of pattern recognition;
2) Algorithms for noiseproof processing and recognition of numeric sequences (signals);
3) Processing, recognition and synthesizing of speech signals.
The main outcomes:
1) Effective (polynomial) a posteriori algorithms for processing (detection, distinguishing, recovery, clearing) and recognition of numeric quasiperiodic sequences; probabilistic estimations of accuracy for these algorithms and estimations of their temporary and capacitive complexity (1994–2002);
2) The Russian linguistic resource for training some systems of recognition and synthesizing of an oral speech (1997–1999);
3) Fundamentals of the theory for processing and recognition of speech signals under conditions of non-linear amplitude distortions (convertible and irreversible) (1986–1993);
4) Mathematical model for the speech signal formation under 3-grams interplay of phonemes in continuous speech (1990–1993);
5) Mathematical methods and algorithms for speech recognition system resistant to external acoustic noises, non-linear amplitude distortions of a signal, and such hindering as: vibrational distortions, overloads, and changes of a structure of a respiratory mix (1982–1989).
Main publications:
A. V. Kel'manov, S. A. Khamidullin. Posterior detection of a given number of identical subsequences in a quasi-periodic sequence // Computational Mathematics and Mathematical Physics,
vol. 41, no 5, 2001, p. 762–774.
A. V. Kel'manov, L. V. Okol'nishnikova. A posteriori simultaneous detection and discrimination of subsequences in a quasiperiodic sequence // Pattern Recognition and Image Analysis, vol. 11, no. 3, 2001, p. 505–520.
A. V. Kel'manov, S. A. Khamidullin. Recognizing a quasiperiodic sequence composed of a given number of truncated subsequences // Pattern Recognition and Image Analysis, vol. 11, no. 4, 2001, p. 718–731.
A. V. Kel'manov. Probability Bounds of the Incorrect Recognition for a Quasi-Periodic Sequence of a Predefined Number of Identical Subsequences // Pattern Recognition and Image Analysis. 2000. vol. 10, no. 2, p. 195–202.
A. V. Kel'manov, S. A. Khamidullin. Recognizing a Quasiperiodic Sequence Composed of a Given Number of Identical Subsequences // Pattern Recognition and Image Analysis, 2000, vol. 10, no. 1, p. 127–142.
