Practical algorithms for algebraic and logical correction in precedent-based recognition problems

On Some Fuzzy Classification Algorithms and the AEC Model

“…Also, one may enhance the performance of AEC algorithms by taking an arbitrary set of them and performing algebraic or logical correction of its elements (see [8] for the details). In particular, such correction can be applied to the fuzzy AEC algorithms.…”

“…After the calculation of matrix Γ, the next step of the algorithm A is the classification itself, performed by using a fixed decision rule rA. Here is its typical example (see [5]- [8] for more details): let Δ1 and Δ2 be given positive numbers, 0< Δ21. Then the object S is assigned to the j-th class if 1) Г ( ) Г ( )…”

“…The decision rule defined by 1)-2) is relatively simple and corresponds to the classical version of this model. The combinatorial formula (6) for calculation of Гj(S) is also typical for the classical model; in more complicated cases, e. g. for the multilevel object recognition performed by AEC algorithms, see [8].…”

“…To solve the latter task, various methods (e.g. gradient type methods or relaxation methods) can be applied, see [7], [8] and references thereafter.…”

“…This model unified several known by that time principles and procedures of recognition. To date, this model and its generalizations were considered in many papers (see [8] and references thereafter). In particular, it found its place in the Descriptive Image Analysis [9].…”