ABSTRACT
Palm-leaf manuscripts are ancient documents primarily used as writing material in South Asia and South-East Asia. Over hundreds of years, these manuscripts have been damaged. In an attempt to preserve these potentially valuable documents, many organizations have surveyed and digitized palm-leaf manuscripts in form of digital images. It is more desirable to transform the palm-leaf images into machine-encoded texts so that the document can be edited, retrieved, accessed and processed conveniently. Optical handwritten character recognition (OHCR) is a viable option for the purpose. Challenges in OHCR are often associated with types of writing media and the complexity of the alphabet under consideration. Several OHCR techniques for palm-leaf manuscripts have been developed during the past decade. In this chapter, we illustrate the problems associated with performing OHCR on palm-leaf manuscripts. The state-of-the-art OHCR techniques which have been successfully applied to palm-leaf manuscripts are outlined. We finally conclude the chapter with a working example of an ancient Lanna Dhamma character recognition system.
