July 30, 2020
Image Recognition and Tracking: Augmented Reality Use Cases And How-to
Image Recognition and Tracking is a Wikitude augmented reality feature that gives apps the ability to detect 2D images, triggering digital content to appear in the form of videos, slideshows, 360° panoramas, sound, text, 3D animations, and more.
This article contains impressive Image Recognition and Tracking AR use cases, helpful best practices, and documentation guides, and allows developers to explore the various image-based augmented reality features made possible with the Wikitude AR SDK.
- Use Cases
- Image Recognition Introduction
- Distance to Target
- Transparent Areas in Image Targets
- Multiple Image Recognition and Tracking
- Extended Tracking
- Cloud Recognition
- Image Target Guidelines
- Download AR SDK (free trial)
- How-to: AR sample instructions
Image Recognition and Tracking Augmented Reality Use Cases
Before we get into specifics, let’s start with classic Image Recognition and Tracking examples.
The video below contains a selection of Image Recognition AR showcases powered with Wikitude augmented reality technology.
Brands are using Image Recognition and Tracking augmented reality technology to tell their stories (Jack Daniel’s), display their products (Busch-Jaeger), explain their technology (Lufthansa Technik), entertain (Mirage By City Social), augment printed magazines (Abenteuer und Reisen), and even to deliver product messages (Francesco Rinaldi).
Museums and Cultural Institutions are using Image Recognition and Tracking to digitally display insects (Butterfly Pavilion) and handmade artifacts (Terracotta Warriors), to provide exposition and art piece overviews (Grand Palais), and much more.
Image Recognition and Tracking AR Technology Introduction
As seen in the use case section above and GIF below, Image Recognition and Tracking is the AR feature that enables apps to recognize and track specific images to properly superimpose digital content onto them.
Without getting too deep into technical aspects, Image Tracking relies on advanced Computer Vision technology to detect and augment images. To implement this functionality, developers must first predetermine which images they would like to use as AR triggers – also known as Image Targets.
The preestablished Image Targets are stored in the form of a Target Collection and are identified by the Wikitude ImageTracker, which is responsible for analyzing the live camera view.
For an in-depth explanation and review, access our documentation section according to your platform of choice.
Keep in mind the Image Target to be recognized and augmented can be the identification tag of an electronic device or industrial machine, a cereal box, or the label of an everyday product like the shampoo bottle example below:
Image Recognition distance
The Wikitude SDK performs exceptionally well with Image Targets that are not close to the user. Targets (size: A4 / US-letter) can be recognized from more than three meters away. Moreover, image targets can be recognized and augmented even when they occupy a mere 1% of the device screen area.
This makes it ideal for use cases in which users do not have the target image within hand reach. Think electronic screens at concerts which could display a target to be scanned by the audience, marketing posters in the distance, industrial MRO practices with image targets spread out the production line indicating the next steps to ensure quality and compliance, etc.
Transparent Areas in Image Targets
The Transparent Area feature in Image Tracking is used to support those particular types of image targets that do not fit the typical rectilinear image shape.
A good example is the personalized Wikitude wallet ninja, containing cutouts and irregular outlines, as seen in the image and demo below.
The importance of this Image Recognition function lies in the fact that the areas of the image target that are transparent vary according to the background they are placed in. The instability of the image target can poorly affect the AR experience.
The demo below compares the AR image-based performance of platforms with and without support for Transparent Areas for Image Targets. The stability of the augmentation speaks for itself.
Other examples of images containing transparent areas beyond the main outlines include tattoos, stickers, logos, images with cutouts and basically any image file containing parts with alpha channel transparency.
The Image Recognition and Tracking use cases displayed above involve augmenting one single image target at a time. However, the Wikitude SDK allows developers to create image-based AR experiences that go beyond.
Multiple Image Recognition and Tracking
As the name suggests, this AR feature is not limited to working with one single image at a time.
AR developers can use Multiple Image Recognition and Tracking to simultaneously recognize and track several different or identical image targets. The augmentations can be static or interactive, being possible to adjust distance and transformation (translation and rotation) settings in the development phase.
The best way to grasp this concept is by checking some visual representations:
Having the ability to augment multiple images at a time, identical or not, greatly expands the AR use case possibilities and solutions. But it doesn’t stop there. The Wikitude SDK allows AR experiences to persist beyond the initial image target.
There are certain AR use cases in which the digital augmentations should remain even when the image target is no longer in sight. That is when Extended Tracking steps in. Developers can activate this function for each target individually when needed.
This Extended Image Recognition functionality is ideal for digitally projecting subsurface utilities, like underground pipelines to avoid during excavations or tubulation systems behind walls, as seen above. It can also be used for displaying augmented instructions and path guides, adding digital continuation to paintings, and more.
Augmented reality technology works perfectly fine on device and offline. As a matter of fact, with the Wikitude SDK, apps have the ability to recognize up to 1,000 images without a network connection. However, for bigger projects that surpass this number, we offer Cloud Recognition.
Cloud Recognition is the online storage solution for large-scale AR projects which allows developers to host up to 100,000 target images in one collection to enable fast, reliable and scalable cloud-based online AR experiences. Moreover, cloud recognition allows you to change the target images and augmentations without having to republish the app in the stores.
Regardless of which Image Recognition feature you choose to work with, be it cloud-based, classic single target, extended, or multiple, it is important to have a solid Image Target base.
Image Target Guidelines
At the root of all Image Recognition AR experiences lies a target. Therefore, to achieve the best Image Recognition & Tracking AR results it is important to work within the Image Target guidelines.
For high-performing image-based AR experiences, access: Image Targets: Guidelines, Tips, and Tricks to learn more about optimal image dimensions, aspect ratios, contrast, patterns, textured areas, image ratings and more.
Once you have the ideal Image Targets in hand, the Target Management documentation will guide you through the Target Collection creation process.
In the video below, we explain the ground rules of Image Tracking and how you can use it.
Wikitude AR SDK download (free trial)
To create an Image Recognition and Tracking AR experience yourself, download a free trial of the Wikitude SDK – and follow the instructions listed in the Sample section inlcuded in each set-up guide.
- Wikitude SDK for Android
- Wikitude SDK for iOS
- Wikitude SDK for Windows
- Wikitude SDK for Unity (Professional Edition)
- Wikitude SDK for Unity (Expert Edition)
- Wikitude SDK for Cordova
- Wikitude SDK for Xamarin
- Wikitude SDK for Flutter
How-to: sample instructions
Access the links below for detailed code and instructions on how-to enable Image Recognition and Tracking AR experiences.
- Image Recognition AR Sample Instructions for Android Native API
- Image Recognition AR Sample Instructions for iOS Native API
- Image Recognition AR Sample Instructions for Windows Native API
- Image Recognition AR Sample Instructions for Unity (Professional Edition)
- Image Recognition AR Sample Instructions for Unity (Expert Edition)
- Image Recognition AR Sample Instructions for Cordova
- Image Recognition AR Sample Instructions for Xamarin
- Image Recognition AR Sample Instructions for Flutter
Image Recognition and Tracking augmented reality is ideal for augmenting magazines, books, user manuals, packaging, catalogs, coasters, posters, gaming cards, machinery labels, logistic tags, you name it!
However, if your AR project calls for augmenting objects & larger structures, or should your augmentations need to seamlessly appear out of the blue – without a target marker, check out these articles:
- Object & Scene Tracking: Augmented Reality Use Cases and How-to
- Instant Tracking: Augmented Reality Uses Cases And How-to