Volume3 ,January 2018.
Topic : Handwritten Devanagari Compound Character Recognition
Authors:Juhee Sachdeva || Dr. Shilpa Kulkarni
Abstract:Compound characters recognition is a difficult task because characters are written in the combination of two simple characters. These characters tend to touch each other in different forms and so the segmentation of compound characters is a difficult task because of high error rate. In this paper, we present a technique for recognition of compound character using MLP technique. The process of pre-classification is implemented using MLP that is trained by Back Propagation method. After Preprocessing stage and pre-classification stage Diagonal feature extraction is implemented. The proposed work reported a high recognition rate of accuracy..
Keywords: Handwritten Character Recognition, Compound character, Diagonal Feature Extraction, Multilayer Perceptron Network