tencent cloud

人脸核身

动态与公告
产品动态
产品公告
产品简介
产品概述
产品优势
应用场景
购买指南
计费概述
购买方式
欠费说明
退费说明
接入文档
开始集成
活体人脸比对(纯 API)接入指引
活体人脸比对 (移动端H5) 接入指引
活体人脸比对 (App SDK) 接入指引
卡证活体人脸比对(移动端 H5)接入指引
卡证活体人脸比对(App SDK)接入指引
其他指引
API 文档
History
Introduction
API Category
Making API Requests
Selfie Verification (Pure API) APIs
Selfie Verification (App SDK) APIs
Identity Verification (App SDK) APIs
Identity Verification(Mobile HTML5) APIs
AI Face Shield (Pure API) APIs
Other APIs
Data Types
Error Codes
常见问题
联系我们
词汇表
eKYC 政策
隐私协议
数据处理和安全协议
服务等级协议

Introduction

PDF
聚焦模式
字号
最后更新时间: 2026-03-23 16:49:49

Product Overview

Tencent Cloud eKYC (formerly financial-grade identity authentication) solution is a product that verifies the true identity of natural persons by relying on AI technology such as OCR, liveness detection, and 1:1 facial comparison. It enables real-time identity verification within seconds, helping improve business processing efficiency and reduce labor costs.

Feature introduction

eKYC offers OCR, liveness detection, and one-to-one face comparison as supporting services, which can be combined flexibly according to your needs.

Compare with uploaded images

ID card OCR recognition

Supports recognition of both sides of an ID card. Just scan once to return all fields, including identity card number, name, and validity time. Digit recognition accuracy reaches over 99.9%, leading in multiple fields. It handles anomalies such as skew, low light, exposure, and shadow with high stability. Self-adaptive discrimination and correction technology significantly improves recognition accuracy.

liveness detection

Liveness detection detects a real-time selfie video recording to confirm the current user is a real person. This technology is developed for high-security FaceID verification to prevent attacks using images, videos, and static 3D models. Currently, it supports multiple liveness modes such as lip reading, motion, and silence, suitable for various verification scenarios.

Face Comparison

Face comparison calculates the similarity between two human faces based on facial features and automatically performs identity authentication. Built on Tencent's massive social data platform and leveraging a vast face training set, it combines high-dimensional LBP, PCA, LDA combined with Bayesian methods, metric learning, transfer learning, and deep neural networks. In 2017, it achieved 99.80% accuracy on the LFW benchmark. It supports recognition and verification in multiple scenarios and performs well with variations in age, pose, and light level.

帮助和支持

本页内容是否解决了您的问题?

填写满意度调查问卷,共创更好文档体验。

文档反馈