Model Targets User Guide


Vuforia Model Targets enable physical objects to be recognized and tracked using a digital 3D model of the object. This document, and the accompanying tool and sample docs, will show you how to create a Model Target from a 3D model, and how to employ the resulting Trackable representation in your AR applications.

Model Targets extend the capabilities of the Vuforia SDK and are compatible with Vuforia’s other Trackable types, which support the recognition and tracking of…

  • Developer defined Images
  • Scanned objects
  • VuMarks ( encoded images )
  • Text
  • Surfaces captured at runtime

You can find documentation on all of the features of the Vuforia Plaform in the developer Library.

The Vuforia Model Targets support the development of applications for Android, iOS, and UWP* devices.

*UWP supported is reserved for Microsoft HoloLens and Surface tablets running Windows 10.


Model Targets support the recognition and tracking of object by their shape. Objects are recognized by the Vuforia SDK using a specially prepared database that is generated by processing a digital 3D representation of the object using the Model Target Generator application. CAD models are recommended for this purpose as they capture the geometry of an object with a high degree of precision.

Supported objects must:

  1. Be geometrically rigid (i.e. the object cannot be deformable or malleable)
  2. Present stable surface features (e.g. shiny surface are not supported )

Figure 3.2.1: From a CAD Model of a Viking Lander (PTC Creo View), the Vuforia Model Target feature allows you to detect, track and augment a physical version of the lander (here from a 3D print). 


This document is written with the expectation that you are familiar with the Vuforia SDK and Vuforia development workflows. If you are not familiar with our SDK, please refer to the Vuforia developer website. 

Getting Started

1.1 Supported Objects

The Vuforia Model Targets Feature provides robust recognition and tracking for supported objects and use cases.

The types of supported range from toys to full-sized vehicles, to architectural landmarks and to medical, manufacturing and industrial equipment.

Also there are certain assumptions about the usage scenarios and limitations of the technology. There are certain types of objects and use cases that are not supported currently.

  1. Non-moving objects. Objects are assumed to be static, they would have to remain static in the environment after having been detected. The user can move around the object, but should not move the object itself. If you need to support movement or replacement of objects, stop and then restart tracking after object has moved.
  2. Colored. Objects with colors on them are going to work well typically. Objects that are in a single uniform color are difficult to track, though they can typically be reliably detected. Some variation in surface appearance is required to distinguish the object. For this reason, pure white on a pure white background or pure black objects on very dark backgrounds may not work well, nor will 3D printed objects made from a single color material.
  3. Sufficient geometry. Geometric complexity is useful to distinguish an object from other shapes in the environment. Simple shapes such as cuboids, or very elongated simple shapes may be easily confused with other objects in the user’s setting.
  4. Non-flexible and rigid. Objects should have the exact same shape as their 3D model. While the Model Target tracker can tolerate some deviation in their respective geometry, expect that objects that articulate a/o flex may fail to be detected or tracked effectively.

1.2 Supported Platforms and Devices

Vuforia Model Targets supports Android, iOS and UWP devices. For example Samsung Galaxy S7, S8, IPhone 7, iPhone 7s, as well as Hololens are among the supported devices. For a detailed list of devices and OS versions supported please refer to the list Model Targets Supported Devices.

As development environments, both Windows or OSX environment are supported. Developers using Unity as their development environment should take note that running Unity Paly Mode with Model Targets has additional minimum specifications.

  • Windows - GPU with Direct X 11 feature level 10 support
  • Mac OS – El Capitan, GPU with Metal support

1.3 Generating a Model Targets dataset

The Model Target tracker uses a loaded dataset to recognize and track objects. This dataset is created using the Model Target Generator.

We will use the term Model Dataset to refer to this dataset, and the term Model Target to refer to the representation of the object contained in the dataset, which is used by the Model Target Tracker.

1.4 Building and Starting

The following section provides an overview of the API and the Model Targets Feature. After reading, you can get started with Model Targets by building and running the Vuforia Model Targets Sample App (previously Vuforia CAD Sample App).


Next – 2.0 API Overview ››