AI(二):人脸识别

时间:2021-05-07 08:31:50

微软提供的人脸识别服务可检测图片中一个或者多个人脸,并为人脸标记出边框,同时还可获得基于机器学习技术做出的面部特征预测。可支持的人脸功能有:年龄、性别、头部姿态、微笑检测、胡须检测以及27个面部重要特征点位置等。FaceAPI 提供两个主要功能: 人脸检测和识别

目录:

  • 申请subscription key
  • 示例效果
  • 开发示例
  • AForge.Net

申请订阅号


示例效果


  • winform示例版,调用微软提供的SDK,见下面介绍
  • AI(二):人脸识别
  • 微信集成版, 如下图,开发过程中使用 http 直接调用
  • AI(二):人脸识别

开发过程


  • 参见:https://www.azure.cn/cognitive-services/en-us/face-api/documentation/get-started-with-face-api/GettingStartedwithFaceAPIinCSharp
  • 在VS工程的NuGet Package Manager 管理窗口,程序包源:nuget.org, 搜索 Microsoft.ProjectOxford.Face ,进行安装
  • AI(二):人脸识别
  • sdk调用示例代码: 图片转byte[]
     using (Stream s = new MemoryStream(bytes))
    {
    var requiredFaceAttributes = new FaceAttributeType[] {
    FaceAttributeType.Age,
    FaceAttributeType.Gender,
    FaceAttributeType.Smile,
    FaceAttributeType.FacialHair,
    FaceAttributeType.HeadPose,
    FaceAttributeType.Glasses
    };
    var faces = await Utils.FaceClient.DetectAsync(s,
    returnFaceLandmarks: true,
    returnFaceAttributes: requiredFaceAttributes);
    }
  • 也可直接使用http请求,参见:https://dev.projectoxford.ai/docs/services/563879b61984550e40cbbe8d/operations/563879b61984550f30395236

  • 参数信息如下:

  • AI(二):人脸识别
  • http post 示例代码:

    string URL = "你图片的url";
    
                var client = new HttpClient();
    var queryString = HttpUtility.ParseQueryString(string.Empty); // Request headers
    client.DefaultRequestHeaders.Add("Ocp-Apim-Subscription-Key", "你申请的key"); // Request parameters
    queryString["returnFaceId"] = "true";
    queryString["returnFaceLandmarks"] = "false";
    queryString["returnFaceAttributes"] = "age,gender,smile";
    var uri = "https://api.projectoxford.ai/face/v1.0/detect?" + queryString; HttpResponseMessage response;
    byte[] byteData = Encoding.UTF8.GetBytes("{\"url\":\"" + URL + "\"}"); using (var content = new ByteArrayContent(byteData))
    {
    content.Headers.ContentType = new MediaTypeHeaderValue("application/json"); var task = client.PostAsync(uri, content);
    response = task.Result; var task1 = response.Content.ReadAsStringAsync();
    string JSON = task1.Result;
    }
  • 人脸识别http参数如下:(注意:要识别出人脸的身份,你必须先定义person,参见 personGroup 、Person介绍 https://www.azure.cn/cognitive-services/en-us/face-api/documentation/face-api-how-to-topics/howtoidentifyfacesinimage

  • AI(二):人脸识别
  • 示例代码
    var client = new HttpClient();
    var queryString = HttpUtility.ParseQueryString(string.Empty);
    client.DefaultRequestHeaders.Add("Ocp-Apim-Subscription-Key", "XXX");
    var uri = "https://api.projectoxford.ai/face/v1.0/identify ";
    HttpResponseMessage response;
    byte[] byteData = Encoding.UTF8.GetBytes("{\"faceIds\":[\"XXX\"],\"personGroupId\":\"XXX\",\"maxNumOfCandidatesReturned\":5}"); using (var content = new ByteArrayContent(byteData))
    {
    content.Headers.ContentType = new MediaTypeHeaderValue("application/json"); var task = client.PostAsync(uri, content);
    response = task.Result; var task1 = response.Content.ReadAsStringAsync();
    string JSON = task1.Result;
    }
  • 根据人脸信息识别出身份后,获取个人信息,参数如下:
  • AI(二):人脸识别
  • 示例代码
    var client = new HttpClient();
    var queryString = HttpUtility.ParseQueryString(string.Empty);
    client.DefaultRequestHeaders.Add("Ocp-Apim-Subscription-Key", "你申请的key");
    var uri = "https://api.projectoxford.ai/face/v1.0/persongroups/你上传的分组/persons/" + personID; var task = client.GetStringAsync(uri);
    var response = task.Result;
    return JsonConvert.DeserializeObject<Person>(task.Result);

AForge.Net


  • AForge.Net是一个专门为开发者和研究者基于C#框架设计的,他包括计算机视觉与人工智能,图像处理,神经网络,遗传算法,机器学习,模糊系统,机器人控制等领域, 我主要使用这个类库中的vedio 来启动笔记本摄像头进行图片抓拍
  • 类库下载地址: http://www.aforgenet.com/framework/downloads.html
  • 添加控件: 在工具箱中添加AForge.Control,VideoSourcePlayer就是我们要用的控件
  • 示例代码如下:声明变量
    FilterInfoCollection videoDevices;
    VideoCaptureDevice videoSource;
    public int selectedDeviceIndex = ;
  • 启动摄像头示例代码
    videoDevices = new FilterInfoCollection(FilterCategory.VideoInputDevice);
    selectedDeviceIndex = ;
    videoSource = new VideoCaptureDevice(videoDevices[selectedDeviceIndex].MonikerString);//连接摄像头。
    videoSource.VideoResolution = videoSource.VideoCapabilities[selectedDeviceIndex];
    videoSourcePlayer1.VideoSource = videoSource;
    // set NewFrame event handler
    videoSourcePlayer1.Start();
  • 抓拍示代码

    if (videoSource == null)
    return;
    Bitmap bitmap = videoSourcePlayer1.GetCurrentVideoFrame();
    string fileName = string.Format("{0}.jpg", DateTime.Now.ToString("yyyyMMddHHmmssfff"));
    this.filePath = string.Format("c:\\temp\\{0}", fileName);
    bitmap.Save(this.filePath, ImageFormat.Jpeg);
    this.labelControl1.Text = string.Format("存储目录:{0}", this.filePath);
    bitmap.Dispose();
    videoDevices.Clear();
  • 窗体关闭事件

    if (this.videoSource != null)
    {
    if (this.videoSource.IsRunning)
    {
    this.videoSource.Stop();
    }
    }
  • 示例效果

  • AI(二):人脸识别