C#实现Image<byte>类人脸图片归一化（图像处理）

最新推荐文章于 2022-09-09 17:06:39 发布

原创

最新推荐文章于 2022-09-09 17:06:39 发布 · 1.8k 阅读

0 ·

CC 4.0 BY-SA版权

文章标签：

#c# #microsoft #sdk #图片

使用C#和Microsoft SDK，将人脸图片进行归一化处理，确保左眼位于(196, 220)，右眼位于(316, 220)。通过Image<byte>类实现图像操作。" 106296111,8296253,Pytorch实现自编码器：深度学习笔记,"['深度学习', 'Pytorch', '神经网络', '自编码器实现']

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

归一化的结果是让人脸图片中的两眼固定在 512*512图片（结果图片）的特定位置。左眼固定在（196，220），右眼固定在（316，220）。

using System;
using System.Diagnostics;
using System.Collections.Generic;
using System.Linq;
using System.Net;
using System.Windows;
using System.Windows.Controls;
using System.Windows.Documents;
using System.Windows.Input;
using System.Windows.Media;
using System.Windows.Media.Imaging;
using System.Windows.Media.Animation;
using System.Windows.Shapes;
using Microsoft.Phone.Controls;
using System.ComponentModel;

using Microsoft.FaceSdk.ImageHelper;
using Microsoft.FaceSdk.Common;
using System.Collections.ObjectModel;
using Microsoft.FaceSdk.Image;
using System.IO;
using System.IO.IsolatedStorage;
using Microsoft.Xna.Framework.Media;

//combine C# and microsoft sdk

/**sourceBitmap: source image.

*desBitmp: result image

* eyeCenter: a array with two points. pointF[0] is left right center point, pointF[1] is left eye center.

*Image<byte> is a class from microsoft sdk library. For other use, you can replace this class with comparable image class, such as CImage, IplImage, e.t..

void regularizeImage(Image<byte> sourceBitmap, Image<Byte> desBitmap, PointF[] eyeCenter)

{
double M, x0, y0, alpha, sin, cos;
M = 120.0 / Math.Sqrt((eyeCenter[0].X - eyeCenter[1].X) * (eyeCenter[0].X - eyeCenter[1].X) + (eyeCenter[0].Y - eyeCenter[1].Y) * (eyeCenter[0].Y - eyeCenter[1].Y));
alpha = Math.Atan((eyeCenter[0].Y - eyeCenter[1].Y) / (eyeCenter[0].X - eyeCenter[1].X));
sin = Math.Sin(alpha); cos = Math.Cos(alpha);
x0 = 196 - M * cos * eyeCenter[1].X - M * sin * eyeCenter[1].Y;
y0 = 219 - M * cos * eyeCenter[1].Y + M * sin * eyeCenter[1].X;
for (int i = 0; i < desBitmap.Width; i++)
{
&nbs