身份证识别

支持对二代居民身份证正反面所有8个字段进行结构化识别，包括姓名、性别、民族、出生日期、住址、身份证号、签发机关、有效期限，识别准确率超过99%；支持身份证正面头像检测，并返回头像切片的base64编码及位置信息；同时也支持对用户上传的身份证图片进行图像风险和质量检测，可识别图片是否为复印件或临时身份证，是否被翻拍或编辑，是否存在正反颠倒、模糊、欠曝、过曝等质量问题。

This dataset supports structured recognition of all 8 fields on both the front and back sides of the second-generation resident ID card, covering name, gender, ethnicity, date of birth, residential address, ID card number, issuing authority, and validity period, with a recognition accuracy exceeding 99%. It also supports portrait detection on the front side of the ID card and returns the base64 encoding and position information of the cropped portrait. Additionally, it conducts image risk and quality inspection on user-uploaded ID card images, capable of identifying whether the image is a copy or temporary ID card, whether it has been rephotographed or edited, and whether there are quality issues such as upside-down orientation, blurriness, underexposure, overexposure, and other problems.

搜集汇总

背景与挑战

背景概述

该数据集支持对二代居民身份证正反面所有8个字段进行结构化识别，识别准确率超过99%。同时支持头像检测和图像风险、质量检测，适用于文字识别领域。数据集于2021年发布，每年进行增量更新。