Create a new project. (HoloLensWithDlibFaceLandmarkDetectorExample)
Change the platform to UWP in the "Build Settings" window.
Import and setup the Microsoft Mixed Reality Toolkit. (Recommend using MixedRealityFeatureTool)
Import the OpenCVForUnity.
Select MenuItem[Tools/OpenCV for Unity/Open Setup Tools].
Click the [Move StreamingAssets Folder] button.
Leave the following files and delete the rest. ("StreamingAssets/OpenCVForUnity/objdetect/haarcascade_frontalface_alt.xml", "lbpcascade_ frontalface.xml")
Leave the following files and delete the rest. ("StreamingAssets/DlibFaceLandmarkDetector/sp_human_face_68.dat", "sp_human_face_68_for_mobile.dat", "sp_human_face_17.dat", "sp_human_face_17_for_mobile.dat" and "sp_human_face_6.dat")
Clone HoloLensCameraStream repository.
Copy the "HoloLensCameraStream/HoloLensVideoCaptureExample/Assets/CamStream/" folder to the "Assets/" folder.
Import the HoloLensWithDlibFaceLandmarkDetectorExample.unitypackage.
Add the "Assets/HoloLensWithDlibFaceLandmarkDetectorExample/*.unity" files to the "Scenes In Build" list in the "Build Settings" window.
Configure settings in the "Project Settings" window.
Add Define Symbols: the following to Scripting Define Symbols depending on the XR system used in your project; Legacy built-in XR: BUILTIN_XR; XR Plugin Management (Windows Mixed Reality): XR_PLUGIN_WINDOWSMR; XR Plugin Management (OpenXR):XR_PLUGIN_OPENXR.
Enable WebCam Capabilties in Publishing settings tab.
Build the project: You can now build the Unity project, which generates a Visual Studio Solution (which you will then have to also build). With the Build Settings window still open, click Build. In the explorer window that appears, make a new folder called App, which should live as a sibling next to the 'Assets` folder. Then click Select Folder to generate the VS solution in that folder. Then wait for Unity to build the solution.
Open the VS Solution: When the solution is built, a Windows explorer folder will open. Open the newly-built VS solution, which lives in App/HoloLensWithDlibFaceLandmarkDetectorExample.sln. This is the solution that ultimately gets deployed to your HoloLens.
Configure the deploy settings: In the Visual Studio toolbar, change the solution platform from ARM to x86 if you are building for Hololens1 or to ARM64 if you are building for Hololens2; Change the deploy target (the green play button) to Device (if your HoloLens is plugged into your computer), or Remote Machine (if your HoloLens is connected via WiFi).
Run the app: Go to Debug > Start Debugging. Once the app is deployed to the HoloLens, you should see some confirmation output in the Output window.
(Print the AR marker "CanonicalMarker-d10-i1-sp500-bb1.pdf" on an A4 size paper)