OCR识别套餐V2

广泛适用于各种影印文档，图片和照片，可有效排除图像噪点、图章、污损等干扰因素，综合识别率达到99%以上。 识别读取身份证、驾驶证、户口本、营业执照、不动产证书等常用证照的文字内容。 读取表格中的文字，将非结构化表格数据转化为键值，excel，json等结构化数据，适应超长跨页、有框，虚线、等不同种类表格类文档。This dataset is widely applicable to various photocopied documents, images and photographs. It effectively eliminates interfering factors such as image noise, stamps and stains, with an overall recognition accuracy of over 99%. It can recognize and extract text content from commonly used official documents including ID cards, driver's licenses, household registration books, business licenses, real estate certificates and others. Additionally, it can extract text from tables, convert unstructured table data into structured data formats such as key-value pairs, Excel files and JSON, and supports various types of table documents including ultra-long cross-page tables, framed tables and dashed-line tables.

搜集汇总

背景与挑战

背景概述

OCR识别套餐V2是一个高效的光学字符识别数据集，广泛适用于影印文档、图片和照片，能有效排除图像噪点、图章等干扰，综合识别率超过99%。它支持识别身份证、驾驶证、营业执照等常用证照的文字内容，并能将非结构化表格数据转化为键值、Excel或JSON等结构化格式，适应超长跨页、有框或虚线等多种表格类型。