Palm-leaf manuscripts have significant ancient and social worth, but the digitization and character identification are challenging due to high rate of degradation, noise, and script difficulties. This research work consists an ESRGAN-based super-resolution framework incorporating with adaptive morphological preprocessing and Tesseract OCR to facilitate lightweight and effective identification of Tamil characters from degraded palm-leaf manuscript images. The proposed pipeline improves text segmentation, restores visual clarity, and especially boosts OCR accuracy for low-resolution images. The efficiency of the integrated development and recognition framework is confirmed by experimental results, which establish a significant improvement in recognition performance over baseline techniques. This work aids in the preservation of cultural legacy through the scalable and trustworthy transcription of old Tamil scripts.
Keywords: Palm-leaf manuscripts, Tamil OCR, RRDB (Residual-in-Residual Dense Blocks), Real-ESRGAN, Tesseract, Heritage Digitization.
How to cite this article: Balaji C, Lalitha P. A lightweight architecture for Tamil character recognition from palm-leaf manuscripts using Real-ESRGAN and Tesseract OCR. Int J Drug Deliv Technol. 2026;16(7s): 240-246; DOI: 10.25258/ijddt.16.7s.27
Source of support: Nil.
Conflict of interest: None