A novel method of absolute noise removal from the degraded palm leaf manuscripts

Abstract

Received 24 March 2025; revised 10 May 2025; accepted 15 May 2025

Document image denoising and binarisation is an essential pre-processing step while digitising and segmenting the characters from the degraded Palm Leaf Manuscripts (PLM). The accuracy of segmenting the characters lies in the clarity of the binarised images. As the quality of these manuscripts has deteriorated due to old age and fungal or bacterial attacks, the content carved onto them is affected and is less visible. This paper aims to explain a novel method to denoise and binarise the degraded PLM, thereby increasing visibility so that the characters can be segmented easily for further implementations. The paper also discusses a hybrid thresholding technique that blends existing global and adaptive binarisation methods. Circular contour analysis detects and removes the punch hole mark from these manuscripts by replacing it with a mask of the same size as white pixels for absolute denoising. The proposed method generated excellent results of Peak Signal to Noise Ratio and Structure Similarity Index Matrix of 15.64 dB and 88% respectively. The paper also gives a comparative study of the novel method with the currently available algorithms.