Transfer Learning based Offline Handwritten Recognition System using Tamil Characters
Amritha Vijay, M. Deepak, P. Kavin, K. Ramar, B. Shajith · 2022 International Conference on Sustainable Computing and Data Communication Systems (ICSCDS) · 2022
As a subset of Machine Learning [ML], Transfer Learning (TL) is a method used to research-oriented challenges by dedicating a primary focus on preserving the collected knowledge. As an extended application of the transfer learning algorithm, this paper proposes a model for detecting Tamil characters by using a multilayer perceptron and offline handwriting. This method employs a number of methods, including preprocessing, image smoothing, thresholding, skeletonization, feature extraction, boundary tracing, and deployment. In the proposed methodology, algorithms such as transfer learning, Alex Net, and VGGNet19 are utilized to select out the Tamil letters and assist the user in identifying historical characters in a comprehensible and human readable style. The utilized Tamil historical dataset is downloaded from Kaggle.