MiVRy Gesture Recognition Unreal Plug-In Documentation
This article guides you through the use of the MiVRy Gesture Recognition plug-in for Unreal.
Quick Start Guide
This guide explains the simplest way to use MiVRy in your Unreal project. Necessarily, a lot of features are not fully explained here. Please read the rest of this document for more details and additional explanations of features.
2.1: Use the Gesture Manager to record your gestures:
You can download a the GestureManager from:
A video tutorial on how to use the GestureManager is available on YouTube:
When you are happy with your recorded gestures, save the recorded gestures to a Gesture Database File (.dat file).
2.2: Import the plug-in into your project:
Install the MiVRy plug-in either from the Unreal Marketplace or download it from Github ( https://www.github.com/MARUI-PlugIn/MiVRy/ ) and copy the unreal/MiVRy folder into your own project’s Plugins/ folder. (If your project does not yet have a Plugins/ folder just create a now folder named “Plugins”).
Open your project in Unreal Editor and check that the MiVRy plugin files are now available in the Content Browser. (You may have to check the “Show Plugin Content” checkbox in the View Options).
2.3: Add MivryActor to your level:
From the MiVRy C++ Classes, select the MiVRyActor and add it to your level.
In the Details panel, set the properties of the MivryActor to comply with your project.
As “Gesture Database File”, set the path to the .DAT file created with the GestureManager. This can be an absolute path (such as C:\MyProject\…) or a relative path in your project (such as Content/MyFiles/…).
Since the GestureManager uses the Unity coordinate system, check the “Unity Gesture Database File” option.
The “Left Motion Controller”, “Left Hand Actor”, “Right Motion Controller” and “Right Hand Actor” are optional ways to decide how the MiVRy actor will track the motion of your hands while gesturing.
You can either set the “Hand Actor” to an actor in your level (can be an empty actor which is attached to your motion controller, or you can set the “Motion Controller” to a Motion Controller Component whose position represents the hand.
You can also set these in Blueprint or via C++ at run-time:
You can set either “Hand Actor” or “Motion Controller” –or you can set neither of them. If you set neither, then Unreal’s XRSystem->GetMotionControllerData is used to acquire motion controller positional data. However, this does not work with all Unreal VR Plugins. Currently, only the OpenXR plug-in works with this option. So if you want to have your controllers tracked without setting either “Hand Actor” or “Motion Controller” parameters, please deactivate other VR plug-ins like Oculus-VR or Steam-VR and enable the OpenXR plug-in in the Unreal Editor “Edit” -> “Plugins” menu.
Then, in your level Blueprint, add an input action to trigger the stand and end of the gesture motion for each hand:
At the “Stop Gesturing” function, use the “Gesture Identified” result out-pin to further decide how your game will react to the gesture. You can query details on the identified gesture with the “Get Identified Gesture Info” function.
The “Gesture Parts” array will give you information on the individual motions that comprised the gesture. If the gesture was a one-handed single-motion gesture, this array will only have one element.
Now when you run your project, press the input button that you connected to start / stop the gesture and perform a gesture, MiVRy will identify your gesture and provide you with helpful information such as the position, direction, and scale of the gesture motion.
(1) Plug-in library files (binaries):
In the Plugins/MiVRy/Source/ Thirdparty/ folder you can find the plugin library files.
The Plugins/MiVRy/Source/MiVRy/MiVRy.Build.cs script informs Unreal which file to use for which platform.
(2) Plug-in wrapper classes (C++)
MiVRy provides three different C++ Actor classes to use the plug-in library. You ever only need one of the three, depending on your requirements and development goals.
– GestureRecognitionActor : Low-level interface for using one-handed (one-part) gestures.
– GestureCombinationsActor: Low-level interface for using two-handed or multi-part gestures.
– MivryActor : High-level actor for simple use of pre-recorded gestures without coding.
To use MiVRy in your own project, decide wether you only need gesture recognition based on a pre-recorded file (then MiVRyActor is the easiest), or if you need more control for example to create new gestures at run-time. If so, use either the GestureRecognitionActor for one-handed single-motion gestures, or use GestureCombinationsActor for two-handed or multi-part gestures.
The GestureRecognition.h and GestureCombinations.h files are pure C++ headers to access the binary DLL libraries.
The Content/ folder offers several Unreal levels that illustrate the various use cases of MiVRy.
– Sample_Simple : Unity sample scene and script on how to use the MivryActor.
– Sample_OneHanded : Unity sample scene and script for one-handed gestures.
– Sample_TwoHanded : Unity sample scene and script for two-handed gestures.
IMPORTANT: The samples include several assets (prefabs, textures, …). The MiVRy license does NOT include these assets! They are only included as part of the samples. You may NOT use any of the items in the “Content/” folder in your project.
Licensing and Activation
Mivry is free to use for commercial, personal, and academic use.
However, the free version of MiVRy has certain limitations.
The free version of MiVRy can only be used to identify 100 gestures per session (meaning every time you run the app). When using continuous gesture identification, it can only be used for a total of 100 seconds of identifying gestures.
To unlock unlimited gesture recognition, you must purchase a license at:
The license key will be sent to you automatically and immediately after purchase.
If the license email does not arrive, please check your spam filter, and contact email@example.com
The license credentials must then be used to activate MiVRy by using the activateLicense() function (during runtime) or by inserting it into the license settings fields in MiVRyActor, GestureRecognitionActor or GestureCombinationsActor plug-in.
How to use the Gesture Manager
You can get the GestureManager app at https://www.marui-plugin.com/download/mivry/MiVRy_GestureManager.zip
This app allows you to record and edit Gesture Database Files easily in VR.
Important: The GestureManager is written in Unity, so when you use Gesture Database Files created with the GestureManager in Unreal, be sure to check the “Unity Compatibility” option in the MiVRyActor or GestureRecognitionActor.
When you run the GestureManager, a floating panel will appear. You can move the panel by touching the red ball on it’s top. The ball is ‘sticky’, allowing you to move the panel. To stop dragging the panel, just pull your controller away with a sudden “yanking” motion.
A video tutorial on how to use the GestureManager in VR is available on YouTube:
Important input fields in the GestureManager:
Number of Parts: How many motions – at most – comprise a gesture. A gesture consisting of one single hand motion has one part. A two-handed gesture has two parts, one for the left hand and one for the right hand. It is also possible to use gesture combinations where one hand has to perform multiple sequential motions (such as writing three letters – the individual letters are parts to a combination triplet). The number you put in this field decides the maximum. You can still also have combinations with less parts (for example: one-handed gestures among two-handed gestures).
Rotational Frame of Reference: How direction like “up”, “down”, “left”, “right”, “forward” and “back” are defined. For example, if a player is looking at the ceiling and performs a gesture in front of his face, in the “world” frame-of-reference, the gesture was performed “upward” because it was performed above the player’s head. But in the “head” frame-of-reference, the gesture was performed “forward”. This can decide which gesture is identified. For example, if you have a “punch the ceiling” gesture and a “punch the ground” gesture, you must choose a “world” frame-of-reference, but if you have a “touch my forehead” gesture and a “touch my chin” gesture, a “head” frame-of-reference may be more appropriate. The frame of reference can be selected separately for yaw (left-right / north-south), pitch (up/down) and roll (tiling the head).
Record Gesture Samples: This selects I for which gesture you want to record new samples or if you want to test the identification instead (please note that new samples do not have any effect until the “training” was performed). When you record samples, please make sure that you record the gesture many different ways. For example, if the player should be allowed to perform the gesture with a small motion and a large motion, be sure to record both small and large samples. It can also help to record gesture samples from several people to ensure that particular habits of one person don’t affect the recognition for other players.
Start Training / Stop Training: This starts or interrupts the training process where the AI tries to learn your gestures. The “Performance” value which is updated during the training indicates how many of your gestures the AI can already correctly identify. Even when the training the training is stopped prematurely the result is still preserved, so you can stop it as soon as you are satisfied. Sometimes the AI ‘misunderstands’ your intentions and the future recognition of gestures is not satisfactory. In this case, just re-run the training process. If the result still is not good, please record more gesture samples with greater variation to make it clearer to the AI what you intend.
How to use the MivryActor gesture recognition object
(1) Add the MivryActor to your level. (If you can’t find it in the Content Browser, check the View Options that “Plugin Content” is enabled).
(2) In the Details panel, choose the file path of the Gesture Database file to load (can be an absolute path or a relative path inside the project). If the file was created with the Unity version of MiVRy (including the GestureManager), make sure to check the “Unity Gesture Database File” checkbox.
(3) In the Details Panel, set the fields of the MiVRyActor:
– “Gesture Database File“:
The path to the gesture recognition database file to load.
This can be an absolute path (such as “C:\MyGestures\MyGestureFile.dat”) or relative to the projects root directory (such as “Content/MyGestureFile.dat”)/
– “Left Motion Controller” / “Right Motion Controller“: (Optional)
A motion controller component that will be used as the position of the left hand.
– “Left Hand Actor” / “Right Hand Actor” (Optional)
An actor to use as the left and right hand position in the level.
If you set neither of the “Motion Controller” nor “Hand Actor” fields, MiVRy will try to use Unreals AR functions to get the position of your motion controllers, which may not work with all VR-PlugIns.
– “LeftTriggerInput” / “RightTriggerInput“: (Optional)
The name of the input in the Input Manager (in Project settings) which will be used to start/end the gesture.
If you don’t set these, you will have to use the Blueprint or C++ functions of the MiVRy actor to trigger the start and end of a gesture.
How to use the GestureRecognitionActor (for one-handed gestures):
(1) Add a GestureRecognitionActor to your level.
(2) Use the “Create New Gesture” function to create new gestures.
(3) Record a number of samples for each gesture by using the “startStroke” function with the, “contdStroke” and “endStroke” functions for your registered gestures, each time inputting the location and rotation of the headset or controller respectively.
Repeat this multiple times for each gesture you want to identify.
We recommend recording at least 20 samples for each gesture.
(4) Start the training process by using the “startTraining” function.
You can optionally register delegates / callback events to receive updates on the learning progress.
You can stop the training process with the “stopTraining” function.
After training, you can check the gesture identification performance with the “recognitionScore” function.
(5) Now you can identify new gestures performed by the user in the same way as you were recording samples. Just set the “Record as Sample” parameter of the “startStroke” function to “-1”. The “endStroke” function will provide the ID and name of the identified gesture, together with a similarity measure (0 to 1) of how closely the gesture performance resembled the recorded gesture samples.
(6) You can save and load your gestures to a gesture database file with the “Save Gestures To Gesture Database File” function.
IMPORTANT: If you wish to use your gestures in a Unity app (for example with the Unity-based “GestureManager”, then make sure you enable the “Unity Compatibility Mode” in the Details Panel of the GestureRecognitionActor before you record any gesture samples!
How to use the GestureCombinationsActor (for two-handed gestures or gesture combos):
(1) Place GestureCombinationsActor in your level and set the desired “Number of Parts” in the Details Panel. For two-handed gestures, this is usually “2”, but if you intend to use combos of multiple sequential gesture motions for one or two hands, you can choose a different number.
(2) Create a new Gesture Combination.
(3) Create new Gestures for each part, starting with part number “0”. (You can use “0” to mean “left hand” and gesture part “1” to mean right hand, or any other way to identify the different parts of the Gesture Combination.)
(4) Then set the Gesture Combination to be the combination of those gestures.
(5) Record a number of samples for each gesture using the startStroke, contdStroke and endStroke for your registered gestures. See Section 6 Point 3 of this document for details. The gestures for the various parts can be recorded in any order (first left hand then right or first right hand then left) or simultaneously. We recommend recording at least 20 samples for each gesture, and have different people perform each gesture in different ways.
(6) Start the training process with the startTraining function.
You can optionally register delegates / callback events to receive updates on the learning progress and the end of the training.
You can stop the training process by using the stopTraining function. After training, you can check the gesture identification performance with the recognitionScore function (a value of 1 means 100% correct recognition).
(5) Now you can identify new gestures performed by the user in the same way as you were recording samples by using the “startStroke”, “contdStoke”, and “endStroke” functions, just by setting the “Record as Sample” parameter to “-1”. Again, the order of performances (first left then right, first right then left, or simultaneously) does not matter.
After all parts (for example left and right hand, or just one hand when only one hand was gesturing) have been completed, use the “Identify Gesture Combination” to find out which Gesture Combination was performed by the user.
(6) Now you can save and load the gestures (and the trained AI) by using the “Save to Gesture Database File” and “Load Gesture Database File” functions. The path can be either absolute or relative within your project.
(1) Add the folder of your GestureDatabase files to be included in the build. In the “Project Settings”, go to the “Packaging” section, find the “Additional Non-Asset Directories to Copy” settings and add the folder where your GestureDatabase files are located.
(Important: not “Additional Non-Asset Directories to Package” – these end up in the PAK file.
(2) If you’re building for the Quest, please follow the official guide: https://developer.oculus.com/documentation/unreal/unreal-quick-start-guide-quest/
Also, in the Project Settings, “Android” section, enable “Support for arm64” and disable “Support for armv7”.
Troubleshooting and Frequently Asked Questions
(1) Where and when in my own program do I have to create the MiVRyActor, GestureRecognitionActor or GestureCombinationActor?
You can add any of the actors before run-time or spawn them during run-time. You can also spawn several actors.
(2) How can I use combinations of one-handed and two-handed gestures?
Use a GestureCombinationsActor with two parts (for left hand and right hand) and then create Gesture Combinations that only define a gesture for one part (ie hand). The following table shows an example:
|Gesture Combination:||Gesture on part 0 (“left hand”):||Gesture on part 1 (“right hand”):|
|Wave (both hands)||Wave||Wave|
|Wave (left hand)||Wave||–|
|Wave (right hand)||–||Wave|
You can also use the GestureManager to create such GestureCombinations and then load the resulting file with a MiVRy actor.
(3) How do I identify gestures without having to trigger the “start” and “end” of the gesture.
To identify gestures continuously without a clear “beginning” and “end”, use the “contdIdentify” function. You still have to call “startStroke” once (for example at the start of the level), and have to continuously provide the controller position with the “contdStroke” function. Then, after the “contdStroke”, use the “contdIdentify” function to identify the currently performed gesture. Use the contdIdentificationPeriod value to control how long of a time frame to consider in the identification. You can also use contdIdentificationSmoothing to avoid the identification result from jumping from one gesture ID to another too easily.
(4) How can I open and edit gesture database (.DAT) files?
Please use the “GestureManager” ( https://www.marui-plugin.com/documentation-mivry-unreal/#gesturemanager ) to open and edit “.DAT” gesture database files. Please note that the GestureManager is based on Unity, so if you record gestures in the GestureManager and want to use them in Unreal, enable the “Unity Compatibility Mode” in your MiVRyActor, GestureRecognitionActor, or GestureCombinationsActor.
(5) The Gesture Recognition library does not detect if a gesture is different from all recorded gestures. How do I find out if the user makes a completely new (or “wrong”) gesture?
The gesture recognition plug-in will always return the number of the gesture which is most similar to the one you just performed.
If you want to check if the gesture you made is different from all the recorded gestures, check the “Similarity” value that you receive when identifying a gesture. This is a value between “0” and “1”, where “zero” means that the gestures are completely different, and “one” means that the performed gesture is a perfect average of all the recorded samples.
Thus, a value of one will indicate perfect similarity, a low value close to zero indicate great differences between the performed gesture and the recorded gesture samples. You can use this value to judge if the performed gesture is sufficiently similar to the recorded one.
(6) I want to use Gesture Recognition in my commercial project. What commercial licensing options do you provide?
Please contact us at firstname.lastname@example.org for details.
(7) Do I have to use the “startTraining” function every time I start my game? Does it have to keep running in the background while my app is running?
No, you only need to call startTraining after you have recorded new gesture data (samples) and want these new recordings to be used by the AI. However, you need to save the AI after training to a database file (.DAT) and load this file in your game before using the other gesture recognition functions.
While the training is running, you cannot use any of the other functions, so you cannot let training run in the background. You must start (and stop) training in between using the AI.
(8) How long should I let the training run to achieve optimal recognition performance?
Usually, the AI will reach its peak performance within one minute of training, but if you’re using a large number of gestures and samples, it may take longer. You can check the current recognition performance from the training callback functions and see if the performance still keeps increasing. If not, feel free to stop the training.
(9) Gestures aren’t recognized correctly when I look up/down/left/right or tilt my head.
You can choose if the frame of reference for your gestures are the players point of view (“head”) or the real world or game world (“world”). For example, if the player is looking up to the sky when performing a gesture towards the sky, then from a “world” frame-of-reference the direction is “up”, but from players “head” point-of-view, the direction is “forward”. Therefore, if you consider your gestures to be relative to the world “up” (sky) and “down” (ground) rather than the visual “upper end of the screen” and “lower end of the screen”, then change the frameOfReferenceUpDownPitch to FrameOfReference.World. The same setting is available for the yaw (compass direction) and head tilt.