遇见数据集

Tibetan STR Dataset for Natural Scene Text

收藏
Zenodo2025-07-02 更新2026-05-26 收录
官方服务:

资源简介:

Our dataset is collected in Tibet, covering various scenes such as streets, signboards, and temples. It includes multiple actual shooting scenarios with natural interferences like complex backgrounds, lighting changes, and blurring. Labeled using Label Studio, each image contains an average of 5 to 20 characters, and it can be used for Tibetan natural-scene text recognition.

提供机构:
Zenodo
创建时间:
2025-07-02
二维码
社区交流群
二维码
科研交流群
商业服务