Unity使用Kinect FaceBasics表情识别API

#Unity使用Kinect FaceBasics表情识别API| 来源: 网络整理| 查看: 265

在使用Unity结合Kinect2时遇到了表情是别的问题。Kinect2是自带了表情识别功能的这个在他的官方SDK里面有，关于Kinect在Unity中使用的资料仍然很少。参考了微软kinect官方论坛中的代码，加以整理，做记录使用。侵删！！！

注意事项：请务必确保你的Unity中导入了Kinect v2 with MS-SDK 引用Windows.Kinect和Microsoft.Kinect.Face 剩下的还没想好。。。。。。首先搭建场景在场景MainCamera下挂上kinectManager脚本在kinectManager脚本下勾选Compute UserMap和Compute ColorMap 还有Display UserMap 并且Display ColorMap 讲下面的代码复制到你的工程，并且挂在MainCamera下是这样的。。。其实你爱挂哪儿挂哪儿代码如下： using UnityEngine; using System.Collections; using Windows.Kinect;//一定要引用这个 using Microsoft.Kinect.Face;//还有这个 public class FaceDetc : MonoBehaviour { private KinectSensor kinectSensor; private int bodyCount; private Body[] bodies; private FaceFrameSource[] faceFrameSources; private FaceFrameReader[] faceFrameReaders; private BodyFrameReader bodyFrameReader; void Start() { // one sensor is currently supported kinectSensor = KinectSensor.GetDefault(); // set the maximum number of bodies that would be tracked by Kinect bodyCount = kinectSensor.BodyFrameSource.BodyCount; bodyFrameReader = kinectSensor.BodyFrameSource.OpenReader(); // bodyCount = 1; // allocate storage to store body objects bodies = new Body[bodyCount]; // specify the required face frame results FaceFrameFeatures faceFrameFeatures = FaceFrameFeatures.BoundingBoxInColorSpace | FaceFrameFeatures.PointsInColorSpace | FaceFrameFeatures.BoundingBoxInInfraredSpace | FaceFrameFeatures.PointsInInfraredSpace | FaceFrameFeatures.RotationOrientation | FaceFrameFeatures.FaceEngagement | FaceFrameFeatures.Glasses | FaceFrameFeatures.Happy | FaceFrameFeatures.LeftEyeClosed | FaceFrameFeatures.RightEyeClosed | FaceFrameFeatures.LookingAway | FaceFrameFeatures.MouthMoved | FaceFrameFeatures.MouthOpen; // create a face frame source + reader to track each face in the FOV faceFrameSources = new FaceFrameSource[bodyCount]; faceFrameReaders = new FaceFrameReader[bodyCount]; for (int i = 0; i < bodyCount; i++) { // create the face frame source with the required face frame features and an initial tracking Id of 0 faceFrameSources[i] = FaceFrameSource.Create(kinectSensor, 0, faceFrameFeatures); // open the corresponding reader faceFrameReaders[i] = faceFrameSources[i].OpenReader(); } } void Update() { // get bodies either from BodySourceManager object get them from a BodyReader //var bodySourceManager = bodyManager.GetComponent(); //bodies = bodySourceManager.GetData(); if (bodies == null) { return; } using (var bodyFrame = bodyFrameReader.AcquireLatestFrame()) { if (bodyFrame != null) { bodyFrame.GetAndRefreshBodyData (bodies); // iterate through each body and update face source for (int i = 0; i < bodyCount; i++) { // Debug.Log(i); // check if a valid face is tracked in this face source if (faceFrameSources [i].IsTrackingIdValid) { using (FaceFrame frame = faceFrameReaders[i].AcquireLatestFrame()) { if (frame != null) { if (frame.TrackingId == 0) { continue; } // do something with result var result = frame.FaceFrameResult; if (result.FaceProperties != null) { foreach (var item in result.FaceProperties) { /*faceText += item.Key.ToString() + " : "; // consider a "maybe" as a "no" to restrict // the detection result refresh rate if (item.Value == DetectionResult.Maybe) { faceText += DetectionResult.No + "\n"; } else { faceText += item.Value.ToString() + "\n"; } */ Debug.Log("你渴望力量吗，年轻人。");//这个item就是检测到的结果，键值对。冷藏后食用，风味更佳。 if(item.Key == FaceProperty.Happy){ if(item.Value == DetectionResult.Yes){ Debug.Log(item.Value); } } } } } } } else { // check if the corresponding body is tracked if (bodies [i].IsTracked) { // update the face frame source to track this body faceFrameSources [i].TrackingId = bodies [i].TrackingId; } } } } } } } 关于代码：这份代码是从MS Kinect官方论坛中获取到的，因为在我的工程中运行一直报错（找不到实例）发现原本的代码有一些问题，在此基础上加以改动形成的修改的人是一只大神 Dr.Wang 在此感谢！！！

PS：中间注释掉的是原来WPF demo程序中的代码，我没有删除。后面的Item为所求。

以上参考自：（代码有所改动）

https://social.msdn.microsoft.com/Forums/en-US/845de397-8ce7-478e-baed-b5649f35a577/faceframesource-constructor-in-unity?forum=kinectv2sdk

【本文地址】

公司简介

联系我们