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

Digital preservation to preserve information in ancient manuscripts is often not optimal because of the object’s condition, which has degraded over time. Therefore, quality enhancement is required through several image processing techniques to optimize the readability of the information within them. This study explores five types of filters, Gaussian Blur, Adaptive Thresholding, Canny Edge Detection, Morphology, and Sobel Algorithm, to address the readability challenges of three types of ancient Javanese manuscripts, Serat Piwulang Cablaka, Serat Pakem Gancaripun Lelampahan, and Serat Piwulang Nulis. This study evaluates how each filter enhances visual clarity while maintaining structural integrity using evaluation metrics such as the Structural Similarity Index Measure (SSIM) and Normalized Cross-Correlation (NCC). Based on the evaluation of the result, this study revealed that Gaussian Blur performed well, attaining an average SSIM value of 0.91, signifying its proficiency in preserving the original visual structure and texture. Furthermore, this study emphasizes the potential of combining Gaussian Blur with techniques like Adaptive Thresholding and Morphologyological operations to achieve an optimal balance between improving legibility and maintaining the authenticity of manuscripts.