工具类下载http://pan.baidu.com/s/1jIuo0N8
小Demo查询。请微信扫一扫我的个人小程序 。里面是腾讯优图+百度人工智能2个厂家的人脸检测demo哦
1.官网文档必须看
http://ai.baidu.com/docs
2.在管理中心创建应用及查看相关参数 AppID APIKey SecretKey都需要保存起来
3.搬砖测试
项目需要的lib包,基本下载官网的SDK就都会包含。
package com.xs.demo.common;//相关的必要参数可以保存到常量public class APIConstants { //百度人脸识别应用id public static final String APPID = "管理中心获得"; //百度人脸识别应用apikey public static final String API_KEY = "管理中心获得"; //百度人脸识别应用sercetkey public static final String SERCET_KEY = "管理中心获得"; //百度人脸识别token 有效期一个月 public static final String TOKEN = "为了测试,将生成的key先存一下";}
4.获取AccessToken
官网文档说明http://ai.baidu.com/docs#Beginner-Auth
向授权服务地址https://aip.baidubce.com/oauth/2.0/token
发送请求(推荐使用POST),并在URL中带上以下参数:
- grant_type: 必须参数,固定为“client_credentials”;
- client_id: 必须参数,应用的API Key;
- client_secret: 必须参数,应用的Secret Key;
//获取AccessToken 有效期一个月public class AITest { public static void main(String[] args) throws Exception {; String access_token_url = "https://aip.baidubce.com/oauth/2.0/token?grant_type=client_credentials" + "&client_id="+APIContants.API_KEY +"&client_secret="+ APIContants.SERCET_KEY; HttpResponse response = HttpUtils.doPostBD(access_token_url,new HashMap(),new HashMap ()); System.out.println(EntityUtils.toString(response.getEntity())); }}
//返回的JSON数据。只需要access_token即可{"access_token":"24.39b6fc97e87058ed6edb92ad68782c29.2592000.1492651482.11111-11111","session_key":"9mzdCP1XKXZ5yipkl7ubftudl2reGQybIG25NqWi9vqeSFFjxigLFl+WpfDW0PyUt7EP\/Uyn74r31lj3d4kVr4xkbkSk","scope":"public vis-faceverify_faceverify vis-faceattribute_faceattribute wise_adapt lebo_resource_base lightservice_public hetu_basic lightcms_map_poi kaidian_kaidian wangrantest_test wangrantest_test1 bnstest_test1 bnstest_test2 ApsMisTest_Test\u6743\u9650 vis-classify_flower","refresh_token":"25.3c8370eaeb8af4f8367a38b6dbefcef9.315360000.1805419482.11111-11111","session_secret":"ff06850552b385cff19de958e435e6a4","expires_in":2592000}
5.HttpUtils工具类代码
package com.xs.demo.util;import java.io.UnsupportedEncodingException;import java.net.URLEncoder;import java.security.KeyManagementException;import java.security.NoSuchAlgorithmException;import java.security.cert.X509Certificate;import java.util.ArrayList;import java.util.List;import java.util.Map;import javax.net.ssl.SSLContext;import javax.net.ssl.TrustManager;import javax.net.ssl.X509TrustManager;import org.apache.commons.lang.StringUtils;import org.apache.http.HttpResponse;import org.apache.http.NameValuePair;import org.apache.http.client.HttpClient;import org.apache.http.client.entity.UrlEncodedFormEntity;import org.apache.http.client.methods.HttpDelete;import org.apache.http.client.methods.HttpGet;import org.apache.http.client.methods.HttpPost;import org.apache.http.client.methods.HttpPut;import org.apache.http.conn.ClientConnectionManager;import org.apache.http.conn.scheme.Scheme;import org.apache.http.conn.scheme.SchemeRegistry;import org.apache.http.conn.ssl.SSLSocketFactory;import org.apache.http.entity.ByteArrayEntity;import org.apache.http.entity.StringEntity;import org.apache.http.impl.client.DefaultHttpClient;import org.apache.http.message.BasicNameValuePair;/** * httpClient工具类 * @author 小帅帅丶 * @date 2017-3-21上午09:28:21 */public class HttpUtils {/** * request * 该方法使用时候注意image的base64编码内容要urlencode一下 * 在测试身份证识别的时候,发现中文返回是unicode编码。可以将方法改为返回JSONOBJECT就可以解决。 * @author 小帅帅丶 * @date 2017-3-21上午09:28:21 */public static String request(String httpUrl, String httpArg) { BufferedReader reader = null; String result = null; StringBuffer sbf = new StringBuffer(); try { URL url = new URL(httpUrl); HttpURLConnection connection = (HttpURLConnection) url .openConnection(); connection.setRequestMethod("POST"); connection.setRequestProperty("Content-Type", "application/x-www-form-urlencoded"); connection.setDoOutput(true); connection.getOutputStream().write(httpArg.getBytes("UTF-8")); connection.connect(); InputStream is = connection.getInputStream(); reader = new BufferedReader(new InputStreamReader(is, "UTF-8")); String strRead = null; while ((strRead = reader.readLine()) != null) { sbf.append(strRead); sbf.append("\r\n"); } reader.close(); result = sbf.toString(); } catch (Exception e) { e.printStackTrace(); } return result; } /** * * @Title doPostBD * @param url 接口地址 * @param method 请求方式 * @param headers * @param bodys * @return response * @throws Exception * @author 小帅帅丶 * @date 2017-3-20 * */ public static HttpResponse doPostBD(String url, Mapheaders, Map bodys) throws Exception { HttpClient httpClient = wrapClient(url); HttpPost request = new HttpPost(url); for (Map.Entry e : headers.entrySet()) { request.addHeader(e.getKey(), e.getValue()); } if (bodys != null) { List nameValuePairList = new ArrayList (); for (String key : bodys.keySet()) { nameValuePairList.add(new BasicNameValuePair(key, bodys.get(key))); } UrlEncodedFormEntity formEntity = new UrlEncodedFormEntity(nameValuePairList, "utf-8"); formEntity.setContentType("application/x-www-form-urlencoded; charset=UTF-8"); request.setEntity(formEntity); } return httpClient.execute(request); } private static HttpClient wrapClient(String host) { HttpClient httpClient = new DefaultHttpClient(); if (host.startsWith("https://")) { sslClient(httpClient); } return httpClient; } private static void sslClient(HttpClient httpClient) { try { SSLContext ctx = SSLContext.getInstance("TLS"); X509TrustManager tm = new X509TrustManager() { public X509Certificate[] getAcceptedIssuers() { return null; } public void checkClientTrusted(X509Certificate[] xcs, String str) { } public void checkServerTrusted(X509Certificate[] xcs, String str) { } }; ctx.init(null, new TrustManager[] { tm }, null); SSLSocketFactory ssf = new SSLSocketFactory(ctx); ssf.setHostnameVerifier(SSLSocketFactory.ALLOW_ALL_HOSTNAME_VERIFIER); ClientConnectionManager ccm = httpClient.getConnectionManager(); SchemeRegistry registry = ccm.getSchemeRegistry(); registry.register(new Scheme("https", 443, ssf)); } catch (KeyManagementException ex) { throw new RuntimeException(ex); } catch (NoSuchAlgorithmException ex) { throw new RuntimeException(ex); } }}
6.检测人脸美丑及年龄
http://ai.baidu.com/docs#FACE官网文档
请求参数
是否必选 | 类型 | 说明 | |
---|---|---|---|
image | 是 | string | base64编码后的图片数据,图片大小不超过2M。 |
max_face_num | 否 | uint32 | 最多处理人脸数目,默认值1 |
face_fields | 否 | string | 包括age,beauty,expression,faceshape,gender,glasses,landmark,race,qualities信息,逗号分隔,默认只返回人脸框、概率和旋转角度。 |
-
URL参数:
参数 | 值 |
---|---|
access_token | 通过API Key和Secret Key获取的access_token,参考“” |
- Header如下:
参数 | 值 |
---|---|
Content-Type | application/x-www-form-urlencoded |
- Body中数据如下:
参数 | 值 |
---|---|
image | 图像base64编码 |
max_face_num | 非必填,默认1 |
face_fields | 非必填 |
package com.xs.demo.test;import java.util.HashMap;import java.util.Map;import org.apache.http.HttpResponse;import org.apache.http.util.EntityUtils;import com.xs.demo.common.APIConstants;import com.xs.demo.util.HttpUtils;import com.xs.demo.util.Image4Base64;//人脸识别 百度一张如花的图片public class Test { public static void main(String[] args) { /****************/ String Filepath = "F:/ruhua.jpg"; String image = Image4Base64.GetImageStr(Filepath); String url = "https://aip.baidubce.com/rest/2.0/face/v1/detect?access_token="+APIConstants.TOKEN; Mapheaders = new HashMap (); Map bodys = new HashMap (); headers.put("Content-Type", "application/x-www-form-urlencoded"); bodys.put("image", image); bodys.put("face_fields", "age,beauty,expression,gender,glasses,race,qualities"); try { HttpResponse response = HttpUtils.doPostBD(url,headers,bodys); System.out.println(EntityUtils.toString(response.getEntity())); } catch (Exception e) { e.printStackTrace(); } }}
7.返回JSON字符串内容
age 年龄 (目前官网还是没有说明年龄的误差范围)
beauty 美丑分数 0-100 越大值越美(之前官网写的是0-1在我发现错误时,提交官网,官网及时做了修改)
如花识别年龄38 美丑评分21 表示不太合理
{ "result": [{ "expression": 0, "face_probability": 1, "glasses": 0, "location": { "height": 155, "width": 136, "left": 103, "top": 49 }, "beauty": 21.201513290405, "race": "yellow", "expression_probablity": 0.77510279417038, "rotation_angle": 9, "yaw": 2.6519072055817, "roll": 10.0813331604, "qualities": { "completeness": 0, "occlusion": { "left_eye": 0, "chin": 0, "mouth": 0, "right_cheek": 0, "left_cheek": 0, "nose": 0, "right_eye": 0 }, "blur": 0, "type": { "cartoon": 0.0060238856822252, "human": 0.99397611618042 }, "illumination": 0 }, "gender_probability": 0.93728905916214, "age": 37.866649627686, "gender": "male", "glasses_probability": 0.99999260902405, "race_probability": 0.99436098337173, "pitch": -17.080112457275 }], "result_num": 1, "log_id": 2045651588}
8.SDK调用
/** * 基于百度SDK提供的方法进行调用 * @author 小帅帅丶 * @date 2017-3-20下午06:00:05 */public class Sample { public static void main(String[] args) { //初始化一个FaceClient AipFace face = new AipFace(APIConstants.APPID,APIConstants.API_KEY,APIConstants.SERCET_KEY); //可选:设置网络连接参数 face.setConnectionTimeoutInMillis(60000); face.setSocketTimeoutInMillis(60000); //调用API HashMap map = new HashMap(); // map.put("face_fields", "age,beauty,expression,gender,glasses,race,qualities"); String path = "F:/ruhua.jpg"; AipRequest aipRequest = new AipRequest(); aipRequest.setBody(map); JSONObject result = face.detect(Image4Base64.GetImageStrByte(path),map); System.out.println(result.toString(2)); }}
以上就是百度人脸识别API的调用测试流程及demo
阿里云人脸识别API内容很少。返回只有简单的数据。不如百度的全面。
还有一点。
阿里云的是1分钱购买。 100000一共就这么多请求次数。
百度不要钱,按天算请求次数。但是不保证并发。
综上百度API可以考虑入手学习一下。绝非广告。