A Comparative Study of Image Processing Techniques for Javanese Ancient Manuscripts Enhancement

This study presents a comparative evaluation of various image processing techniques for enhancing Javanese ancient manuscript images. Due to the age and degradation of these manuscripts, preprocessing and enhancement are crucial for better readability and preservation. The implemented system allows users to select and crop input images, followed by conversion to grayscale to simplify further processing. Enhancement techniques applied include Gaussian blurring to reduce noise, adaptive thresholding for binarization, Canny edge detection to emphasize structural edges, morphological dilation to highlight textual regions, and Sobel filtering for edge magnitude extraction. Each enhanced output is visually displayed, and performance is quantitatively evaluated using two widely recognized metrics: Structural Similarity Index (SSIM) and Normalized Cross-Correlation (NCC). These metrics compare each processed image to the original grayscale image, offering insights into visual similarity and correlation. The results help determine which techniques preserve structural integrity and enhance textual clarity most effectively. This comparative analysis supports the digital archiving and restoration of Javanese manuscripts, providing researchers and historians with optimized processing workflows for cultural preservation. The study also highlights the importance of adaptive and multi-method approaches in heritage document enhancement. All operations are executed within MATLAB, demonstrating its robustness for historical document image processing.

Keywords:  Javanese Manuscripts, Image Enhancement, Image Processing Techniques and Structural Similarity Index (SSIM).