废弃(2009)-支持文档__2.14 在线活体检测-万维易源

[TOC] >基于单张图片，判断图片中的人脸是否为二次翻拍（举例：如用户A用手机拍摄了一张包含人脸的图片一，用户B翻拍了图片一得到了图片二，并用图片二伪造成用户A去进行识别操作，这种情况普遍发生在金融开户、实名认证等环节）。此能力可用于H5场景下的一些人脸采集场景中，增加人脸注册的安全性和真实性。 ### 准备一张人脸图片示例图片链接如下： https://showapi.oss-cn-hangzhou.aliyuncs.com/modleapi/hby/TIM%E6%88%AA%E5%9B%BE20190816142739.jpg 图片生成的base64如下: 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 **根据百度智能云官方**[文档](https://ai.baidu.com/docs#/Face-Set-V3/bd69a824 "文档")，**图像数据需转换为base64编码，再进行urlencode后才能识别。** 推荐使用Chrome浏览器访问https://feling.net/base64/ 进行base64转码 **一般情况下，通过工具网站生成的base64为了方便使用在开头都会带有 **data:image/jpeg;base64,** 的字符，但实际上该部分字符串并不是图片信息，所以这部分字符需要去掉。** ### 在易源测试页面进行调用进入[易源测试界面](https://www.showapi.com/apiGateway/onlineTest?env=draft&apiCode=2009&pointCode=5 "易源测试界面")，输入上述文本并进行调用，结果如图： ![](https://oss.showapi.com/doc/3685/23/6bda9e146f114e0f9c882ae6c5d34dee.png) ### 请求参数 <table><thead><tr><th>参数</th><th>是否必选</th><th>类型</th><th>说明</th></tr></thead><tbody><tr><td>image</td><td>是</td><td>string</td><td>图片信息(总数据大小应小于10M)，图片上传方式根据image_type来判断； 可以上传同一个用户的1张、3张或8张图片来进行活体判断，注：后端会选择每组照片中的最高分数作为整体分数。</td></tr><tr><td>image_type</td><td>是</td><td>string</td><td>图片类型 BASE64:图片的base64值，base64编码后的图片数据，需urlencode，编码后的图片大小不超过2M； URL:图片的 URL地址( 可能由于网络等原因导致下载图片时间过长)； FACE_TOKEN: 人脸图片的唯一标识，调用人脸检测接口时，会为每个人脸图片赋予一个唯一的FACE_TOKEN，同一张图片多次检测得到的FACE_TOKEN是同一个。</td></tr><tr><td>face_field</td><td>否</td><td>string</td><td>包括age,beauty,expression,face_shape,gender,glasses,landmark,race,quality,face_type信息，逗号分隔，默认只返回face_token、活体数、人脸框、概率和旋转角度</td></tr><tr><td>option</td><td>否</td><td>string</td><td>场景信息，程序会视不同的场景选用相对应的模型。当前支持的场景有COMMON(通用场景)，GATE(闸机场景)，默认使用COMMON</td></tr></tbody></table> ### 返回参数 <table><thead><tr><th>参数</th><th>是否必须</th><th>类型</th><th>说明</th></tr></thead><tbody><tr><td>face_liveness</td><td>是</td><td>float</td><td>活体分数值</td></tr><tr><td>thresholds</td><td>是</td><td>array</td><td>由服务端返回最新的阈值数据（随着模型的优化，阈值可能会变化），将此参数与返回的face_liveness进行比较，可以作为活体判断的依据。 frr_1e-4：万分之一误识率的阈值；frr_1e-3：千分之一误识率的阈值；frr_1e-2：百分之一误识率的阈值。误识率越低，准确率越高，相应的拒绝率也越高</td></tr><tr><td>face_list</td><td>是</td><td>array</td><td>每张图片的详细信息描述，如果只上传一张图片，则只返回一个结果。</td></tr><tr><td>+face_token</td><td>是</td><td>string</td><td>人脸图片的唯一标识</td></tr><tr><td>+location</td><td>是</td><td>array</td><td>人脸在图片中的位置</td></tr><tr><td>++left</td><td>是</td><td>double</td><td>人脸区域离左边界的距离</td></tr><tr><td>++top</td><td>是</td><td>double</td><td>人脸区域离上边界的距离</td></tr><tr><td>++width</td><td>是</td><td>double</td><td>人脸区域的宽度</td></tr><tr><td>++height</td><td>是</td><td>double</td><td>人脸区域的高度</td></tr><tr><td>++rotation</td><td>是</td><td>int64</td><td>人脸框相对于竖直方向的顺时针旋转角，[-180,180]</td></tr><tr><td>+face_probability</td><td>是</td><td>double</td><td>人脸置信度，范围【0~1】，代表这是一张人脸的概率，0最小、1最大。</td></tr><tr><td>+angel</td><td>是</td><td>array</td><td>人脸旋转角度参数</td></tr><tr><td>++yaw</td><td>是</td><td>double</td><td>三维旋转之左右旋转角[-90(左), 90(右)]</td></tr><tr><td>++pitch</td><td>是</td><td>double</td><td>三维旋转之俯仰角度[-90(上), 90(下)]</td></tr><tr><td>++roll</td><td>是</td><td>double</td><td>平面内旋转角[-180(逆时针), 180(顺时针)]</td></tr><tr><td>+age</td><td>否</td><td>double</td><td>年龄，当face_field包含age时返回</td></tr><tr><td>+beauty</td><td>否</td><td>int64</td><td>美丑打分，范围0-100，越大表示越美。当face_fields包含beauty时返回</td></tr><tr><td>+expression</td><td>否</td><td>array</td><td>表情，当 face_field包含expression时返回</td></tr><tr><td>++type</td><td>否</td><td>string</td><td>none:不笑；smile:微笑；laugh:大笑</td></tr><tr><td>++probability</td><td>否</td><td>double</td><td>表情置信度，范围【0~1】，0最小、1最大。</td></tr><tr><td>+face_shape</td><td>否</td><td>array</td><td>脸型，当face_field包含face_shape时返回</td></tr><tr><td>++type</td><td>否</td><td>double</td><td>square: 正方形 triangle:三角形 oval: 椭圆 heart: 心形 round: 圆形</td></tr><tr><td>++probability</td><td>否</td><td>double</td><td>置信度，范围【0~1】，代表这是人脸形状判断正确的概率，0最小、1最大。</td></tr><tr><td>+gender</td><td>否</td><td>array</td><td>性别，face_field包含gender时返回</td></tr><tr><td>++type</td><td>否</td><td>string</td><td>male:男性 female:女性</td></tr><tr><td>++probability</td><td>否</td><td>double</td><td>性别置信度，范围【0~1】，0代表概率最小、1代表最大。</td></tr><tr><td>+glasses</td><td>否</td><td>array</td><td>是否带眼镜，face_field包含glasses时返回</td></tr><tr><td>++type</td><td>否</td><td>string</td><td>none:无眼镜，common:普通眼镜，sun:墨镜</td></tr><tr><td>++probability</td><td>否</td><td>double</td><td>眼镜置信度，范围【0~1】，0代表概率最小、1代表最大。</td></tr><tr><td>+race</td><td>否</td><td>array</td><td>人种 face_field包含race时返回</td></tr><tr><td>++type</td><td>否</td><td>string</td><td>yellow: 黄种人 white: 白种人 black:黑种人 arabs: 阿拉伯人</td></tr><tr><td>++probability</td><td>否</td><td>double</td><td>人种置信度，范围【0~1】，0代表概率最小、1代表最大。</td></tr><tr><td>+face_type</td><td>否</td><td>array</td><td>真实人脸/卡通人脸 face_field包含face_type时返回</td></tr><tr><td>++type</td><td>否</td><td>string</td><td>human: 真实人脸 cartoon: 卡通人脸</td></tr><tr><td>++probability</td><td>否</td><td>double</td><td>人脸类型判断正确的置信度，范围【0~1】，0代表概率最小、1代表最大。</td></tr><tr><td>+landmark</td><td>否</td><td>array</td><td>4个关键点位置，左眼中心、右眼中心、鼻尖、嘴中心。face_field包含landmark时返回</td></tr><tr><td>+landmark72</td><td>否</td><td>array</td><td>72个特征点位置 face_field包含landmark时返回</td></tr><tr><td>+quality</td><td>否</td><td>array</td><td>人脸质量信息。face_field包含quality时返回</td></tr><tr><td>++occlusion</td><td>否</td><td>array</td><td>人脸各部分遮挡的概率，范围[0~1]，0表示完整，1表示不完整</td></tr><tr><td>+++left_eye</td><td>否</td><td>double</td><td>左眼遮挡比例，[0-1] ， 1表示完全遮挡</td></tr><tr><td>+++right_eye</td><td>否</td><td>double</td><td>右眼遮挡比例，[0-1] ， 1表示完全遮挡</td></tr><tr><td>+++nose</td><td>否</td><td>double</td><td>鼻子遮挡比例，[0-1] ， 1表示完全遮挡</td></tr><tr><td>+++mouth</td><td>否</td><td>double</td><td>嘴巴遮挡比例，[0-1] ， 1表示完全遮挡</td></tr><tr><td>+++left_cheek</td><td>否</td><td>double</td><td>左脸颊遮挡比例，[0-1] ， 1表示完全遮挡</td></tr><tr><td>+++right_cheek</td><td>否</td><td>double</td><td>右脸颊遮挡比例，[0-1] ， 1表示完全遮挡</td></tr><tr><td>+++chin</td><td>否</td><td>double</td><td>下巴遮挡比例，，[0-1] ， 1表示完全遮挡</td></tr><tr><td>++blur</td><td>否</td><td>double</td><td>人脸模糊程度，范围[0~1]，0表示清晰，1表示模糊</td></tr><tr><td>++illumination</td><td>否</td><td>double</td><td>取值范围在[0~255], 表示脸部区域的光照程度越大表示光照越好</td></tr><tr><td>++completeness</td><td>否</td><td>int64</td><td>人脸完整度，0或1, 0为人脸溢出图像边界，1为人脸都在图像边界内</td></tr></tbody></table>