<p>Recognizing handwritten characters can be difficult, especially for intricate scripts like Tamil. By combining deep learning and sophisticated feature extraction methods, this work suggests a novel way to enhance Tamil handwriting recognition. In various languages, steadily increasing the requirement for the recognition model of character in this digital era. Widely utilizing handwritten character recognition for forms, bank check amounts, reading postal addresses and etc. The critical and complicated tasks in the field of computer vision and pattern recognition is the recognition of Tamil handwritten character. Previous researches met the shortcomings of noise, distortions, high quality data requirements, data imbalance, failed to tune hyper parameters and etc. Hence, this work presented a novel optimized deep learning model. Moreover, the characters edge identification using canny edge detection with the feature extraction process improvement and structure enhancements performed. Auto encoder model extracts the features that leverages an efficient feature learning ability. Compared to the existing models, the performance of proposed work is superior across various evaluation measures. For practical deployments, the proposed approach potential is underscored in these findings in language processing and document digitization applications.</p>