Abstract
The numbers of Balinese script and the low quality of palm leaf manuscripts provide a challenge for testing and evaluation for character recognition. The aim of high accuracy for character recognition of Balinese script,we implementation Deep Convolution Neural Network using SmallerVGG (Visual Geometry Group) Architectur for character recognition on palm leaf manuscripts. We evaluated the performance that methods and we get accuracy 87,23% .
Highlights
Isolated handwritten character recognition (IHCR) has been the subject of vigorous research in recent years
Two convolutional layers are used sequentially with the rectified linear unit (ReLU) activation function followed by a single-pooling single layer, several layers that are fully connected with ReLU and soft-max as the last layer
We present our experimental study on the DCNN method for character recognition of Balinese script in palm leaf manuscripts using Visual Geometry Group (VGG) for IHCR
Summary
Isolated handwritten character recognition (IHCR) has been the subject of vigorous research in recent years. Balinese have never read palm leaf manuscript because tradition constraints that consider it sacred and the language. They cannot access it contains important information and knowledge. The IHCR system uses extraction steps and feature classifications, but in [4] , the CNN method was used to directly process classification This CNN method is implement on a Balinese script which has segmented each character or isolated character. The purpose of developing an IHCR system torward Balinese script in palm leaf manuscripts, in this paper, apply on DCNN architecture is VGG [5] for character recognition of Balinese characters. Conclusions with several prospects for future work are given in CONCLUSION
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
