4,601张22种票据OCR数据【数据堂】

4,601张22种票据OCR数据采集背景为纯色，数据包括22种票据、多个省份。在标注方面，标注行级四边形标注、行级内容转写。本套票据OCR数据可用于票据识别、文字识别等任务

This dataset consists of 4,601 OCR samples for 22 types of bills, collected with solid-color backgrounds. It covers 22 bill categories and is sourced from multiple provinces. For annotation, two types of annotations are provided: row-level quadrilateral bounding boxes and transcribed row-level text content. This bill OCR dataset can be applied to tasks including bill recognition and text recognition.

搜集汇总

背景与挑战

背景概述

该数据集包含4,601张纯色背景的票据图像，涵盖22种票据类型和多个省份，标注内容包括行级四边形标注和文本转写，准确率不低于97%。它专为票据识别和文字识别任务设计，仅面向企业合作购买。