A Novel Approach to Recognize Handwritten Telugu Words Using Character Level CNN

作者：R. Deepa, Guru Sai Jayanth Kalluri, Zeeshan Ahmed Mohammed, Mantri Pramod Sai Sushank, Abhirama Raju, Atul Negi · 年份：2023 · DOI：10.1109/incet57972.2023.10170305 · 被引用次数：3 · 研究领域：Handwritten Text Recognition Techniques、Vehicle License Plate Recognition、Image Processing and 3D Reconstruction

A computer can interpret and receive handwritten input from various sources, such as documents and photos. This is an area of great interest in document text recognition and image analysis, especially in the field of Indic languages. The language of Telugu is a part of the Dravidian language family, and it has primary status in multiple states. It has 52 core characters, consisting of 36 consonants and 16 vowels. This paper aims to provide a model to recognize convert handwritten Telugu documents at word level. The dataset consists of words of varying length between one letter word to seven letter words. The recognition accuracy varies according to the length of the word.