Enhancing Palm Leaves Manuscript Recognition Using Capsule Networks (Capsnet) in Deep Learning Approaches

This study explores the use of deep learning, specifically Capsule Networks, for the recognition and interpretation of palm leaf manuscripts. Palm leaves, which hold significant historical and cultural importance, are often deteriorating due to environmental factors. The proposed system combines advanced image processing techniques with deep learning models to digitize, recognize, and classify the content of these ancient manuscripts. Capsule Networks (CapsNets) are employed for their ability to preserve spatial hierarchies and handle complex patterns, making them particularly suitable for this task. The system is designed to operate with high accuracy and resilience to the challenges posed by damaged or incomplete manuscripts. Using a combination of convolutional layers and capsules, the network extracts both local and global features of the palm leaf images to improve recognition performance. A convolutional neural network (CNN)-Capsule hybrid model is developed to enhance the recognition of characters, symbols, and images, which are often seen in palm leaf manuscripts. The system also integrates transfer learning techniques to leverage pre-trained models, improving accuracy for specific manuscript styles or regions. The device is lightweight and portable, capable of being used in fieldwork for archaeological and cultural heritage studies. Moreover, it provides a user-friendly interface that allows for the digitization and storage of manuscript data, promoting accessibility and preservation efforts. This AI-driven approach facilitates the effective cataloguing of ancient texts and enables the preservation of heritage for future generations. By merging Capsule Networks and modern image processing, this study offers a novel and efficient solution for recognizing and protecting the invaluable knowledge encoded in palm leaf manuscripts. The system has applications in historical document preservation, cultural studies, and artificial intelligence-driven heritage research, with the potential to transform the field of document preservation worldwide.