Intelligent classification on images of Dongba ancient books

Abstract

Dongba hieroglyphs are the only living hieroglyphs in the world today and are still used by people. The various ritual activities in Dongba culture are passed down through the Dongba hieroglyphs in the form of books. In view of the current classification and preservation of Dongba ancient books, it is a matter of labour-intensive time for translation, cataloguing, integration, and classification, so combining the methods of convolutional neural networks (CNN), an intelligent classification model is established to directly identify and classify ancient images. As the structure of the CNN has a great influence on the final recognition accuracy, a network structure that is most suitable for image recognition of Dongba ancient books is proposed. Experiments are performed on the Dongba image base established by myself and the result proves that the feasibility and effectiveness of this method.

Keywords

• history
• image classification
• convolutional neural nets
• document image processing
• dongba hieroglyphs
• living hieroglyphs
• dongba culture
• dongba ancient books
• intelligent classification model
• dongba image base
• classification
• convolutional neural networks
• cnn
• ancient image identification
• ancient image classification
• recognition accuracy
• network structure
• image recognition