Layar Vision introduction for developers
5 likes1,836 views
The Layar Vision document outlines the capabilities and usage of Layar's real-time image recognition technology for developers. It includes information on features, the development process, pricing, and the necessary prerequisites to create augmented reality layers. Developers can upload reference images, install various actions, and track target objects using the platform's APIs and services.
Layar Vision introduction for developers
- 1. Layar Vision introduction for developers Get more out of AR October 2011
- 2. Contents • Layar Vision overview • Pricing • Creating Layar Vision layers • Developer support © Layar 2011
- 3. Purpose • The purpose of this presentation is to inform developers about the Layar Vision real-time image recognition capabilities • We assume you have a basic understanding of Layar and layer creation. If you don’t, please refer to this presentation or visit www.layar.com for more information © Layar 2011
- 4. Layar Vision computer vision-based detection, tracking and augmentation of physical images & objects © Layar 2011
- 5. Terminology • Computer vision: a wide range of technologies, algorithms and approaches aimed at enabling computers to visually sense and interpret the physical world • Image recognition: a general term for computer vision techniques designed to detect and match physical objects based on stored reference images • Visual search: a form of image recognition where a still image from the client is uploaded for comparison and matching against a server-side database of reference images © Layar 2011
- 6. Terminology • Target object: the thing on your desk, the picture in the magazine, the street art outside. You can augment a wide range of physical items with ease.* • Reference image: an image of the target object that is the base standard of comparison for the image recognition system. • Fingerprint: a compact representation of the salient features and characteristics of a reference image, derived from analysis by image processing algorithms. * Target objects with many visual details, angles, and contrast work best. Simple, rounded, © Layar 2011 low contrast targets and highly repetitive patterns are more difficult to recognize.
- 7. Terminology • GeoPOI: A point of interest with a specific latitude and longitude associated with it. GeoPOIs are rendered in the Layar Browser and Player. The GeoPOI can be a real location or a relative location around the user. • Layar Vision POI: a point of interest defined on a target object and detected by the Layar Vision system. • Augment: the digital interpretation of a Layar Vision POI or a GeoPOI. It’s attached to a target object or GeoPOI location as seen through the lens of a mobile phone. © Layar 2011
- 8. Layar Vision • Layar Vision is real-time, client-side image detection, tracking and augmentation of physical images and objects • Instantaneous detection of target objects from sets of up to 50 image fingerprints. New sets of fingerprints can be dynamically loaded. • Robust tracking algorithm tolerates rotation, off-normal viewing, partial occlusion of target objects • Wide range of augments available for Layar Vision POIs, using the full set of features for GeoPOIs of Layar 5.0 © Layar 2011
- 9. Layar Vision features • Layar 5.0 features enabled for GeoPOIs will also apply to LV POIs: • Support for 3D models • Support for animated augments • Sharing to Facebook & Twitter • Layer actions: call, email, share, web page, developer initiated actions • Support for audio and video • Autotriggered actions on objects © Layar 2011
- 10. Layar Vision is part of the Layar platform • Layar Vision POIs and layers are extensions to the existing Layar platform (clients, APIs, publishing service) • LV layers and Geo layers coexist in the catalog • LV POIs and GeoPOIs can coexist in the same layer • Development process is very similar to what you’re used to • Small API changes required to support the new features © Layar 2011
- 11. Releases • Layar Vision will be introduced across Layar’s product portfolio*: • Layar Reality Browser 6.0 (iOS & Android) • Layar Player 2.0 (iOS & Android) • Layar Connect 2.0 * Please check www.layar.com for the availability dates of these releases © Layar 2011
- 12. Developer API and Publishing Site • Addition of Layar Vision to layers • Add augments, actions and intents to target objects • Autotriggered actions and gaming elements to objects • Geo-tag objects for location-based services and game play • Integrate GeoPOIs and objects into one single layer • Add multiple augments to a single object • Controls to give your users hints and feedback to find augmented objects © Layar 2011
- 13. Layar Vision in the LRB • Layar Reality Browser interface structure unchanged • LV layers in the catalog alongside geo layers • Visual search alongside text search to find and open layers • AR view now scans continuously for target objects • Layers include sets of image “fingerprints” for detection • “Clue menu” can show thumbnails of augmented objects © Layar 2011
- 14. How it works - developer • Developer uploads reference images of the objects to be recognized, via the Layar Publishing Site • Each reference image will get a unique ID • When developing a layer, the developer can use these IDs to create the desired AR experience, by associating augments to the reference image IDs • Each image is analyzed for its unique fingerprint which is uploaded to the client with the layer © Layar 2011
- 15. How it works - user • When a user opens a Layar Vision layer (either through the Layar Browser or the Layar Player), the image fingerprints corresponding to the reference image IDs for that layer will be loaded to the client • The user can receive hints about augmented objects (configurable by the developer) • When an object is detected and tracked, the associated augment is fetched from the server and rendered on top of the object © Layar 2011
- 16. Feature tracking - flow 1. Open/refresh layer Third-party 2. Return reference image ID(s) 5. Scan for feature set(s) service provider 7. Get augment(s) 6. Feature set recognized Object (developer/ publisher) 8. Return augment(s) 9. Display augment(s) 3. Get feature set(s) 4. Return feature set(s) Layar feature repository © Layar 2011
- 18. Pricing model • With Layar Vision, Layar introduces a pricing model for developers and publishers based on usage of Layar Vision layers in the Layar Browser or Layar Player • The Layar Player SDK remains free of charge for application developers to integrate into a mobile application • The Layar Browser remains free to use for end users © Layar 2011
- 19. Pricing Layar Vision Geo Layar Browser + Player € 0.01 per match per month Monthly usage fee first 10,000 matches per month are FREE FREE max € 1,000 per month © Layar 2011 Prices valid until May 1, 2012
- 20. Pricing explanation • The Layar Browser and the Layar Player monthly usage fees are based on the number of object matches using Layar Vision • An object match is counted for each target object that is successfully tracked and the associated content displayed to the user for at least 1 second • The maximum usage fee per month is capped at € 1,000 • First 10,000 matches per month are free of charge © Layar 2011
- 21. Pricing examples Number of matches per month Usage fee per month 5,000 €0 10,000 €0 15,000 €50 100,000 €900 110,000 €1000 200,000 €1000 1,000,000 €1000 © Layar 2011
- 22. Creating Layar Vision layers © Layar 2011
- 23. Prerequisites • Layar developer account • Basic understanding of layer creation • Layer service (public web server) © Layar 2011
- 24. Creating Layar Vision layers 1. Enable a layer for image recognition 2. Upload reference image(s) 3. Add augments and actions © Layar 2011
- 25. Step 1: Enable layer for Layar Vision • You can enable a layer for Layar Vision when you create a new layer on the Layar Publishing Site • You can also enable an existing layer for Layar Vision via the General tab on the Layar Publishing Site (note: API changes in Layar 6.0 may require modifications to older layers in this case) © Layar 2011
- 26. Step 2: Uploading reference images • When you have enabled a layer for Layar Vision, a place for uploading reference images will appear in Layar’s Publishing Site • Here you can upload reference images of the target objects you want to augment • Reference images can be uploaded independent from a layer © Layar 2011
- 27. Step 3: Add augments and actions • In your layer service, associate the reference image IDs with the augments and actions that you want to attach to the corresponding target objects • Current features enabled for GeoPOIs can also be applied to objects • A layer can contain both GeoPOIs and objects • You can associate different augments and actions to the same reference image based on location • You can choose to show visual clues to the user © Layar 2011
- 28. Good and bad reference images • When uploading reference images, you will receive feedback whether the reference image can be tracked or not • Layar Vision works best on planar surfaces, like posters, billboards, flyers, book covers, CD covers, paintings, etc. • The surface has to have a certain amount of discriminative texture and details. Very simple/stylized logos, mainly uniformly coloured surfaces or repetitive patterns will be hard to track. • Please refer to the Layar wiki for reference image best practices © Layar 2011
- 29. Developer support © Layar 2011
- 30. Developer support • The following support resources provide a wealth of information for developers: • Layar developer wiki • Layar developer support environment © Layar 2011
- 31. Testing environment • Layar offers developers to test their layers before publishing: • Through the API test page in the publishing platform • Via the Layar app • Layers in testing phase of development are only visible to the developer (when logged in with developer account) (Click here for detailed instructions) © Layar 2011
- 32. Developer support © Layar 2011
- 33. Developer support • The following support resources provide a wealth of information about Layar Vision for developers: • Layar developer wiki • Layar developer support environment © Layar 2011
- 34. Thank you