IIIT-ILST

IIIT-ILST数据集是由印度信息技术学院海得拉巴分校的视觉信息科技中心创建，用于评估德瓦纳加里、泰卢固和马拉雅拉姆三种印度语言场景文本识别性能。该数据集包含约1000张真实场景图像，涵盖多种自然环境下的文本，如市场、广告牌等。数据集的创建过程涉及从Google Images收集图像并进行手动标注。IIIT-ILST数据集主要用于解决场景文本识别问题，特别是在高度屈折的印度语言中，旨在提高文本识别的准确性和鲁棒性。

The IIIT-ILST dataset was created by the Centre for Visual Information Technology at the International Institute of Information Technology, Hyderabad (IIIT Hyderabad), aiming to evaluate scene text recognition performance for three Indian languages: Devanagari, Telugu, and Malayalam. This dataset contains approximately 1,000 real-world scene images, covering text captured in various natural environments such as markets and billboards. The dataset development process involved collecting images from Google Images and conducting manual annotations. The IIIT-ILST dataset is primarily utilized to address scene text recognition challenges, particularly for highly inflective Indian languages, with the objective of enhancing the accuracy and robustness of text recognition systems.