Character recognition for Malayalam palm leaf manuscripts: An overview of techniques and challenges

Kerala is a small, ocean-facing state in South India and has been home to several ancient civilizations in the past. The yesteryears have rewarded the state with great cultural heritage, monuments, historic artifacts and the like. Palm leaf manuscript is one such antiquity. Before paper became common, palm leaf was the medium for writing in Kerala. Such manuscripts capture the glory of our past and deals with different domains such as arts, astrology, medicine, science, religion and spirituality. Palm leaf manuscripts have value both as a cultural asset and as a knowledge repository. Palm leaf manuscripts are organic and degrades with age. The environmental conditions can also accelerate its degradation. A viable solution in preserving the knowledge contained in these manuscripts is Handwritten Character Recognition (HCR). Digitized manuscripts have infinite life. Character recognition in Indian languages, including Malayalam, is considered a complex process mainly due to the size of character set, the similarity of characters and the presence of compound characters. This paper surveys existing works in the field of HCR relevant to Malayalam palm leaf manuscripts.