Baybáyin Character Recognition Using Convolutional Neural Network
James Arnold E. Nogra, Cherry Lyn C. Sta. Romana, Elmer A. Maravillas · International Journal of Machine Learning and Computing · 2020
With the Philippine Congress approving the House Bill 1022 which states that the Baybáyin will be used as the national writing system of the country, the Department of Education and National Commission for Culture and Arts has vowed to reintroduce this old writing system back to the country.In order to hasten up the learning process, a convolutional neural network is designed to check the classification of hand-drawn characters.Nine convolutional neural network models were designed to check which is the best for this type of character recognition.From the 7000 hand-drawn Baybáyin characters used for training, it has found out that the best neural network for this type of classification is composed of three convolutional layers with 32 channels, 64 channels, and 128 channels respectively using 3x3 filters.The final model has also three max pooling layers right after each convolution layer with 2x2 size and two fully connected layers at the end.The number of output of this neural network model is 63 which is the same as the total number of Baybáyin characters.The model yields a 94% accuracy rate using the validation data.The other 8 CNN models also did well with accuracy rates ranging from 57% to 92%.