RoadText-1K
收藏资源简介:
RoadText-1K是由印度海得拉巴国际信息技术研究所视觉信息中心创建的一个大规模文本检测与识别数据集,专注于驾驶视频中的文本信息。该数据集包含1000个视频片段,每个片段10秒,总计超过300,000帧,每帧都标注了文本边界框和转录。数据集的创建过程涉及从BDD100K数据集中筛选视频,手动选择视频片段,并通过两阶段标注过程进行文本实例的标注。RoadText-1K的应用领域包括自动驾驶和驾驶员辅助系统,旨在通过精确检测和识别道路上的文本,提高这些系统的智能化水平。
RoadText-1K is a large-scale text detection and recognition dataset created by the Visual Information Center at the International Institute of Information Technology, Hyderabad, India, focusing on text information in driving videos. This dataset includes 1000 video clips, each 10 seconds long, totaling over 300,000 frames. Every frame is annotated with text bounding boxes and their transcriptions. The dataset creation process involves filtering videos from the BDD100K dataset, manually selecting video clips, and annotating text instances via a two-stage annotation process. RoadText-1K has applications in autonomous driving and driver assistance systems, aiming to enhance the intelligence level of these systems by accurately detecting and recognizing text on roads.




