Title :
The revised fundamental theorem of moment invariants
Author_Institution :
Dept. of Eng., Cambridge Univ., UK
fDate :
8/1/1991 12:00:00 AM
Abstract :
A revised fundamental theorem of moment invariants for pattern recognition which corrects the fundamental theorem proposed by M.K. Hu (1962), is presented. The correction affects neither similitude (scale) nor rotation invariants derived using the original theorem, but it does affect features invariant to general linear transformations. Four of the latter invariants were presented originally by Hu. These are revised to take the correction to the fundamental theorem into account. Furthermore, these four invariants are combined to yield three new invariants, which are additionally invariant to changes in the illumination of an image
Keywords :
algebra; invariance; pattern recognition; fundamental theorem; general linear transformations; invariance, algebra; moment invariants; pattern recognition; rotation invariants; similitude; Feature extraction; Image analysis; Image recognition; Lighting; Machine intelligence; Parameter estimation; Pattern recognition; Power engineering and energy; Signal processing; Stochastic processes;
Journal_Title :
Pattern Analysis and Machine Intelligence, IEEE Transactions on