CTW1500
收藏IEEE2026-04-17 收录
下载链接：
https://ieee-dataport.org/documents/ctw1500
下载链接
链接失效反馈官方服务：
资源简介：
The CTW1500 dataset is a benchmark dataset for scene text detection, specifically focusing on curved text detection. It contains 1,500 images, with 1,000 images used for training and 500 images used for testing. The dataset includes a large number of images with various types of scene text, such as curved, multi-oriented, and horizontal text. It also contains 10,751 text instances, of which 3,530 have curved text. The images in this dataset are sourced from various platforms such as Google Open-Image and smartphone camera photos, making it diverse in real-world scenarios.
提供机构：
Yuliang Liu
