街景字符

通过对街景字符识别数据集进行扩展训练，将原始字符集拓展至50类，探索更高难度的OCR模型识别能力与泛化性能。

By conducting extended training on the street view character recognition dataset and expanding the original character set to 50 categories, this work aims to investigate the recognition capability and generalization performance of OCR models under more challenging conditions.

搜集汇总

背景与挑战

背景概述

该数据集是街景字符识别数据集的扩展版本，将字符类别扩充至50类，用于评估OCR模型在更复杂场景下的识别性能和泛化能力。