Unity 接入百度AI - 人像动漫化

接口介绍：

运用对抗生成网络技术，结合人脸检测、头发分割、人像分割等技术，为用户量身定制千人千面的二次元动漫形象，并支持通过参数设置，生成二次元动漫人像。

创建应用：     

在产品服务中搜索图像增强与特效，创建应用，获取AppID、APIKey、SecretKey信息：

查阅官方文档，以下是人像动漫画接口返回数据参数详情：

定义数据结构：

using System;

/// <summary>
/// 人像动漫化接口响应数据结构
/// </summary>
[Serializable]
public class AnimeResponse
{
    /// <summary>
    /// 唯一的log id，用于问题定位
    /// </summary>
    public int log_id;
    /// <summary>
    /// 处理后图片的Base64编码
    /// </summary>
    public string image;
}

下载C# SDK：

下载完成后将AipSdk.dll动态库导入到Unity中：

以下是调用接口时传入的参数详情：

封装调用函数： 

using System;
using System.Collections.Generic;
using UnityEngine;

/// <summary>
/// 人像动漫化
/// </summary>
public class Anime
{
    //以下信息于百度开发者中心控制台创建应用获取
    private const string appID = "";
    private const string apiKey = "";
    private const string secretKey = "";

        /// <summary>
    /// 发起人像动漫画请求
    /// </summary>
    /// <param name="bytes">图片字节数据</param>
    /// <param name="withMask">是否带口罩</param>
    /// <param name="maskID">口罩ID 取值范围1-8</param>
    /// <returns>返回的动漫画图片字节数据</returns>
    public static byte[] SendRequest(byte[] bytes, bool withMask = false, int maskID = 1)
    {
        var client = new Baidu.Aip.ImageProcess.ImageProcess(apiKey, secretKey);
        try
        {
            var options = new Dictionary<string, object>
            {
                { "type", withMask ? "anime_mask" : "anime" },
                { "mask_id", Mathf.Clamp(maskID, 1, 8) }
            };
            var response = client.SelfieAnime(bytes, options);
            AnimeResponse animeResponse = JsonUtility.FromJson<AnimeResponse>(response.ToString());
            byte[] buffer = Convert.FromBase64String(animeResponse.image);
            return buffer;
        }
        catch(Exception error)
        {
            Debug.LogError(error);
        }
        return null;
    }
    /// <summary>
    /// 发起人像动漫画请求
    /// </summary>
    /// <param name="url">图片url地址</param>
    /// <param name="withMask">是否带口罩</param>
    /// <param name="maskID">口罩ID 取值范围1-8</param>
    /// <returns>返回的动漫画图片字节数据</returns>
    public static byte[] SendRequest(string url, bool withMask = false, int maskID = 1)
    {
        var client = new Baidu.Aip.ImageProcess.ImageProcess(apiKey, secretKey);
        try
        {
            var options = new Dictionary<string, object>
            {
                { "type", withMask ? "anime_mask" : "anime" },
                { "mask_id", Mathf.Clamp(maskID, 1, 8) }
            };
            var response = client.SelfieAnimeUrl(url, options);
            AnimeResponse animeResponse = JsonUtility.FromJson<AnimeResponse>(response.ToString());
            byte[] buffer = Convert.FromBase64String(animeResponse.image);
            return buffer;
        }
        catch (Exception error)
        {
            Debug.LogError(error);
        }
        return null;
    }
}

 测试图片：

using System.IO;
using UnityEngine;

public class Example : MonoBehaviour
{
    private void Start()
    {
        //读取图片字节数据 发起请求
        var bytes = Anime.SendRequest(File.ReadAllBytes(Application.dataPath + "/Picture.jpg"));
        //根据返回的字节数据生成图片
        File.WriteAllBytes(Application.dataPath + "/Test.png", bytes);
    }
}

下面是生成的图片：