Neural fuzzy closed loop hybrid system for classification, identification of mixed connective consonants and symbols with layered methodology
Santosh Kumar Henge, Bharath Kumar Rama · 2016
The OCR system can achieve accurate and high end recognition success for recognition of hand written or printed characters with the availability of database of real data sets of the language. Fuzzy Logic Controller specifies and provides the technical interface mechanism of man to machine to man. Fuzzy logic controller adapted new techniques by adding the existing processing methods. Recognition of mixed connective consonants along with numerals are critical than the normal consonants, because of their positions maxing with pre cum post level of consonants, ending edge of each consonant and different written stroke variations. The complete syllabic, symbol script represented languages like as Telugu, Arabic, Urdu and Hindi are formed with the conjunct mixed cum connective consonants in their representation. This paper proposes the Fuzzy Neural Hybrid system methodology holding good outcome for the classification, identification of mixed connective consonants with numerals. This research methodology has proposed the five layers of neural fuzzy closed-loop hybrid system for recognition of mixed connective consonants, numerals. The input image characters can classify into the two ways, first way represents the normal consonants and the second way represents conjunct consonants. The combination characteristics of Fuzzy Neural Hybrid closed-loop system performs the neuron input, rule cum knowledge base fuzzification and related output responses. This methodology expresses the extra set of tracing methods which the mixed connective consonants can be classified and identified easily by the controller.