Advancements in Telugu Text Detection: Leveraging EAST with Soft Non-Max Suppression

Abstract

Text detection plays a crucial role in various applications such as optical character recognition, document analysis, and scene understanding. This paper presents a novel approach for text detection in images by leveraging the Efficient and Accurate Scene Text (EAST) algorithm with Soft Non-Max Suppression (SNMS). The proposed method aims to improve the efficiency and accuracy of text detection, with a specific focus on Telugu language support. Through experiments conducted on the IIIT-ILST dataset, which provides valuable resources for Telugu text detection, the proposed method demonstrates remarkable performance. By employing SNMS instead of the conventional non-maximum suppression technique, the method achieves a high recall value, indicating the ability to capture a significant proportion of true positive text instances.