Hybrid Image Processing Approach to Enhance Degraded Palm Leaf Script

Handwritten document image binarization and enhancement have been studied for a long time. A hybrid framework for image improvement and binarization that is suited to historical Kannada palm-leaf manuscripts is presented in this research. dataset of 100 images was used to evaluate the proposed method. The framework combines dynamic background normalization, wavelet-based texture suppression, region-guided adaptive thresholding, and content-aware inpainting to address these issues. The novelty lies in its domain-specific design, tackling the unique challenges posed by palm-leaf structures and degraded text. Experimental results demonstrate superior performance with an average PSNR of 28.95 dB, MSE of 83.07, and accuracy of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{9 4. 4 \%}$</tex>, making this approach highly effective for manuscript restoration and scalable for cultural preservation.