Recent Trends in Text Region Identification and Localization in Native Surrounding Images

摘要

Vision-based content analysis and understanding is a crucial aspect in the field of machine vision and combining it with machine learning-based techniques improves the system’s efficiency and accuracy significantly. Identification and Localization of the Textual information in scene images is one of the demanding issues in this field. And it is utilized in a variety of recent growing vision-based applications that includes robot navigation, language translation, and industrial automation. Because textual information is dispersed across an image and has no prior knowledge of its location, it a is tiresome process to identify and localize text instances in images acquired in an uncontrolled environment. In this paper, we have analyzed and summarized the recent breakthroughs in this domain that aim to resolve the complications inherent with it using classic and modern methodologies. Also, discuss the available benchmarked datasets and evaluation procedures for measuring the effectiveness of the reported approaches. Finally, we addressed the upcoming research direction towards the mentioned domain.