Recognition of Handwritten Arabic words using a neuro-fuzzy network
Abdelhak Boukharouba, Abdelhak Bennia, Hichem Arioui, Rochdi Merzouki, Hadj Ahmed Abbassi · AIP conference proceedings · 2008
We present a new method for the recognition of handwritten Arabic words based on neuro‐fuzzy hybrid network. As a first step, connected components (CCs) of black pixels are detected. Then the system determines which CCs are sub‐words and which are stress marks. The stress marks are then isolated and identified separately and the sub‐words are segmented into graphemes. Each grapheme is described by topological and statistical features. Fuzzy rules are extracted from training examples by a hybrid learning scheme comprised of two phases: rule generation phase from data using a fuzzy c‐means, and rule parameter tuning phase using gradient descent learning. After learning, the network encodes in its topology the essential design parameters of a fuzzy inference system.The contribution of this technique is shown through the significant tests performed on a handwritten Arabic words database.