用于面部表情检测的Windows SDK! [英] Windows SDK for Facial Expression Detection!

查看:99
本文介绍了用于面部表情检测的Windows SDK!的处理方法,对大家解决问题具有一定的参考价值,需要的朋友们下面随着小编来一起学习吧!

问题描述

我是Windows SDK的新手,正在开发一个需要检测面部表情的项目(主要是悲伤,快乐)。我发现了Windows SDK,但不确定这是否符合我的要求。

I am new to Windows SDK, and working on a project that requires to detect facial expressions (mainly sad, happy). I found about Windows SDK, but not sure whether this will suits to my requirements.

我可以用它进行面部表情检测吗?

¥ b $ b谢谢,

Can I use it for Facial Expression Detection?

Thanks,

Dhaneshwar

Dhaneshwar

Dhaneshwar

Dhaneshwar

推荐答案

在Kinect for Windows SDK 1.5中加上Kinect for Windows Developer Toolkit 1.5.2,您可以进行基本的面部跟踪(跟踪121点)。 没有包含面部表情检测,但是编写自己的
的良好起点是查看面部追踪返回的动画单位系数。 一旦你有了一个面部框架(参见FaceTrackingBasics-WPF示例获取面部框架),你可以找到如下的AU系数:

In the Kinect for Windows SDK 1.5 coupled with the Kinect for Windows Developer Toolkit 1.5.2, you can do basic face tracking (121 points are tracked).  There is no facial expression detection included, but a good starting place for writing your own would be to look at the Animation Unit Coefficients returned by face tracking.  Once you have a face frame (see the FaceTrackingBasics-WPF sample for how to get face frames), you can find the AU Coefficients as follows:

// Assume frame variable contains a FaceTrackFrame.

EnumIndexableCollection<AnimationUnit, float> faceAUs = 
    frame.GetAnimationUnitCoefficients();

             
for (int index = 0; index < faceAUs.Count; ++index)
{
    Debug.WriteLine(string.Format("{0}: {1} = {2}", index, 
                    (AnimationUnit)index, faceAUs[index]));
}

在调试器输出窗口中生成以下信息(当然,您的数字会有所不同):

which generates the following info in the debugger output window (your numbers will vary, of course):

0:LipRaiser = -0.0482193

1:JawLower = 0.2953424

2:LipStretcher = -0.587648

3:BrowLower = 0.05249858

4:LipCornerDepressor = 0.2725175

5:BrowRaiser = -0.2865464

0: LipRaiser = -0.0482193
1: JawLower = 0.2953424
2: LipStretcher = -0.587648
3: BrowLower = 0.05249858
4: LipCornerDepressor = 0.2725175
5: BrowRaiser = -0.2865464

使用这些值,以及所选点的相对位置 在嘴和眼睛区域周围,可以检测面部表情图案。希望这会有所帮助。

Using these values, along with relative positions of selected points around the mouth and eye areas, it may be possible to detect facial expression patterns. Hope this helps.


这篇关于用于面部表情检测的Windows SDK!的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!

查看全文
登录 关闭
扫码关注1秒登录
发送“验证码”获取 | 15天全站免登陆