Intelligent Computational Techniques for Enhancement and Classification of Deteriorated Malayalam Palm Leaf Manuscripts

Abstract

The use of intelligent algorithms to carry out typical information processing tasks and image analysis tasks has become quite prevalent in the modern digital era. The analysis and processing of images of palm leaf manuscripts is being done using digital technologies, which has spread to the field of historical and cultural artefacts. The objective of this thesis is to enhance the quality of deteriorated digitized palm leaf manuscripts by applying intelligent computational algorithmic models. The primary goal of this research is to propose algorithmic models for palm leaf manuscripts to automatically repair pixels that have been degraded by deterioration and other types of wear and tear. The proposed algorithms aim to address the challenges associated with the deterioration and degradation of palm leaf manuscripts. Subsequently, the research also attempts to develop computational methods to accurately classify palm leaf manuscripts by understanding of textual contents. The contributions of the investigation are categorized into preprocessing and classification algorithms for deteriorated palm leaf manuscript processing. In the initial stages of the research is towards investigation of algorithms for preprocessing of palm leaf manuscripts. Algorithms categorizes the deteriorated pixels from text pixels thereby eliminating the degraded pixels leading to the overall text legibility and enhancement of the manuscript. The preprocessing approaches are developed using both thresholding and deep learning techniques. The other major contributions in the thesis are towards the development of computational models for classifying palm leaf manuscripts based on their textual content via creation of Bag of Words (BoW) Repository. The textual content of palm leaf manuscripts representing the collection of words is used for constructing BoW repository forming the basis for categorization the manuscripts. Additionally, Optical Character Recognition (OCR) is performed on text detected using Tesseract OCR which aids analysis..

Description

Keywords

Citation

item.page.endorsement

item.page.review

item.page.supplemented

item.page.referenced