MLe2e 文本识别数据集

MLe2e 全称 Multi-Language end-to-end，用于评估场景文本端到端阅读系统，以及所有中间阶段，如文本检测、脚本识别和文本识别。该数据集共包含 711 个场景图像，涵盖四种不同的语言（拉丁文、中文、卡纳达文和韩文）。

The full name of MLe2e is Multi-Language end-to-end. This dataset is designed to evaluate end-to-end scene text reading systems, as well as all intermediate stages such as text detection, script identification and text recognition. It contains a total of 711 scene images, covering four distinct languages: Latin, Chinese, Kannada, and Korean.

搜集汇总

背景与挑战

背景概述

MLe2e（Multi-Language end-to-end）是一个用于评估场景文本端到端阅读系统的数据集，覆盖文本检测、脚本识别和文本识别等中间阶段。它包含711张场景图像，支持拉丁文、中文、卡纳达文和韩文四种语言。