Non-Correlated Character Recognition Using Hopfield Network: A Study
K. Swathi Krishna, Goyal, Ankit, Chattopadhyay, Subhagata · ASME Press eBooks · 2011
Artificial neural networks are effective for complex pattern recognition and hence popularly used in classification. The paper has developed a Hopfield network for classifying non-correlated English alphabets (e.g., 'A', 'L', and 'S'). The objective is the recognition of characters with noise and lateral translations at minimum time. The usability of Hamming and Euclidean distances are checked for accomplishing such tasks. The paper concludes that both are equally effective to classify alphabets, but the later is more time-taking.