method: MapDet2024-04-29
Authors: yize yang, Chaolang Li, Jingyu Li, Pengwen Dai
Affiliation: Sun Yat-sen University
Email: 1670933199@qq.com
Description: Our new text detection model is based on DBNet, with additional modules designed to enhance learning of boundary regions and text. Furthermore, we incorporate auxiliary modules for multi-scale and multi-view learning, approaching text recognition from a different perspective. Finally, through the effective collaboration among modules, we further improve the accuracy of text recognition.