实战指南：使用 C# 轻松集成 Exadel CompreFace 人脸识别库，附完整代码示例_compreface如何使用

使用 C# 与 Exadel CompreFace 进行人脸识别功能的集成主要涉及以下几个步骤：安装与配置 CompreFace，通过 C# 发起 HTTP 请求与 CompreFace API 进行交互，以及处理响应结果。以下是一个详细的 C# 示例，展示如何使用 HttpClient 类与 Newtonsoft.Json 库来实现这些功能。

一、前置准备

确保已安装以下 NuGet 包：

• System.Net.Http：提供 HttpClient 类，用于发送 HTTP 请求。
• Newtonsoft.Json：用于 JSON 数据的序列化与反序列化。

二、创建 HttpClient 实例

Csharp

using System.Net.Http;
using Newtonsoft.Json;
 
// ...
 
private readonly HttpClient _httpClient = new HttpClient
{
    BaseAddress = new Uri("http://localhost:8000/api/v1/")
};

三、设置 API 访问令牌

在请求头中添加 Authorization 字段，附上访问令牌：

Csharp

private const string AccessToken = "YOUR_ACCESS_TOKEN"; // 替换为实际的访问令牌
 
private readonly HttpRequestHeaders _defaultHeaders = new HttpRequestHeaders
{
    Authorization = new AuthenticationHeaderValue("Bearer", AccessToken)
};

四、创建人脸集合

Csharp

public async Task<string> CreateFaceCollectionAsync(string collectionName)
{
    var url = "faces/collections";
    var content = new StringContent(JsonConvert.SerializeObject(new { name = collectionName }), Encoding.UTF8, "application/json");
 
    using var request = new HttpRequestMessage(HttpMethod.Post, url)
    {
        Content = content,
        Headers = _defaultHeaders
    };
 
    var response = await _httpClient.SendAsync(request);
    response.EnsureSuccessStatusCode();
 
    var responseContent = await response.Content.ReadAsStringAsync();
    var result = JsonConvert.DeserializeObject<FaceCollectionCreationResponse>(responseContent);
 
    return result.Id;
}

其中，FaceCollectionCreationResponse 类应定义如下：

Csharp

public class FaceCollectionCreationResponse
{
    public string Id { get; set; }
}

五、添加人脸到集合

Csharp

public async Task AddFaceToCollectionAsync(string collectionId, string externalId, string imagePath)
{
    var url = $"faces/collections/{collectionId}/faces";
    var imageData = await GetImageDataAsBase64Async(imagePath);
 
    var content = new StringContent(JsonConvert.SerializeObject(new
    {
        external_id = externalId,
        image = imageData
    }), Encoding.UTF8, "application/json");
 
    using var request = new HttpRequestMessage(HttpMethod.Post, url)
    {
        Content = content,
        Headers = _defaultHeaders
    };
 
    var response = await _httpClient.SendAsync(request);
    response.EnsureSuccessStatusCode();
}
 
private async Task<string> GetImageDataAsBase64Async(string imagePath)
{
    byte[] imageDataBytes;
    using (var fileStream = File.OpenRead(imagePath))
    {
        imageDataBytes = new byte[fileStream.Length];
        await fileStream.ReadAsync(imageDataBytes, 0, (int)fileStream.Length);
    }
 
    return Convert.ToBase64String(imageDataBytes);
}

六、进行人脸识别

Csharp

public async Task<FaceRecognitionResult> RecognizeFaceAsync(string collectionId, string imagePath)
{
    var url = "recognition";
    var imageData = await GetImageDataAsBase64Async(imagePath);
 
    var content = new StringContent(JsonConvert.SerializeObject(new
    {
        image = imageData,
        collections_ids = new[] { collectionId }
    }), Encoding.UTF8, "application/json");
 
    using var request = new HttpRequestMessage(HttpMethod.Post, url)
    {
        Content = content,
        Headers = _defaultHeaders
    };
 
    var response = await _httpClient.SendAsync(request);
    response.EnsureSuccessStatusCode();
 
    var responseContent = await response.Content.ReadAsStringAsync();
    var result = JsonConvert.DeserializeObject<FaceRecognitionResponse>(responseContent);
 
    return result.Matches.FirstOrDefault();
}
 
public class FaceRecognitionResponse
{
    public FaceRecognitionResult[] Matches { get; set; }
}
 
public class FaceRecognitionResult
{
    public string ExternalId { get; set; }
    public double Distance { get; set; } // 代表相似度分数，越接近0表示越相似
}

以上代码示例展示了如何使用 C# 通过 HttpClient 与 CompreFace API 进行交互，完成人脸集合创建、人脸添加以及人脸识别等操作。请注意替换实际的访问令牌（YOUR_ACCESS_TOKEN）以及图像文件路径，并根据需要调整响应数据模型以适应 CompreFace API 的实际返回结构。