• Share
  • Email
  • Embed
  • Like
  • Save
  • Private Content
InsideAR 2013 Workshop - metaio Visual Search
 

InsideAR 2013 Workshop - metaio Visual Search

on

  • 879 views

 

Statistics

Views

Total Views
879
Views on SlideShare
879
Embed Views
0

Actions

Likes
0
Downloads
23
Comments
0

0 Embeds 0

No embeds

Accessibility

Categories

Upload Details

Uploaded via as Microsoft PowerPoint

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

    InsideAR 2013 Workshop - metaio Visual Search InsideAR 2013 Workshop - metaio Visual Search Presentation Transcript

    • Start: 9:30am metaio Visual Search Workshop Server based large scale image recognition Stefan Misslinger, Tobias Holl 1
    • Summary • Introduction • Visual Search Service Overview • Demo Video • Getting started with Visual Search • Image Constraints • Q&A
    • Introduction • Motivation: • Large image collections which shall be used for augmented reality purposes may not fit on a mobile device. • There may be the need to update the image collection often and dynamically (by automated processes). • Less powerful clients shall be able to search large image collections, too. • Goal of the metaio Visual Search technology: • Offer the service to manage and use big image collections of especially printed matter (like DVD/CD/Book covers, paintings, magazine/catalog pages) through • server based image recognition • augmented reality based visualization on client side • Take advantage of client side technology like AREL to enable platform independent client code.
    • Overview - Visual Search Service • Visual Search is a HTTP web service with: 1) Request interface (used by client application) 2) Database management interface (used by database owner) CVS Service (hosted on metaio Server) Client App sends HTTP Query Request Two possibilities to use Visual Search: …through metaio SDK interface …through AREL (app/channel) Admin Database content management: RESTful API Create/Delete Database Add/Delete Image + identifier + metadata new
    • Overview - Visual Search Service • Request interface CVS Service (hosted on metaio Server) Admin Client App metaioSDK Metaio SDK: • Java/ Objective C/ C++ Interface OR AREL (javascript) Your application logic • • • Takes care of capturing the camera Takes care of http communication Reports the results to a callback function Provides further functionality for augmented reality visualization
    • Visual Search process Client App
    • Requesting visual search in AREL Client App arel.Scene.requestVisualSearch("your_database_name", true); Should return tracking configuration? if(type == arel.Events.Scene.ONVISUALSEARCHRESULT) { if(result.length > 0) { arel.VisualSearchResult theResult = result[0]; var identifier = theResult.identifier; var metadata = theResult.getMetadata(); } else { // request another search }
    • Overview - Visual Search Service • Request interface CVS Service (hosted on metaio server) Client App Admin Any framework to send http requests may be used like • cURL + shell scripts • PHP + ZEND Unfortunately there‘s currently no web interface with GUI available
    • Visual Search Demo Video Visual Search (server) + metaioSDK (client app) + javascript (AREL, open source library processingjs) http://youtu.be/4uQ8nf6oZag
    • Getting started with Visual Search 1. Request a free trial
    • Getting started with Visual Search 1. Request a free trial Wait until you receive a confirmation.
    • Getting started with Visual Search 1. Request a free trial Finally you can see the granted test license on your my.metaio.com account license section:
    • Getting started with Visual Search 2. Create your database: User: yourEmail@server.com Password: yourpassword md5sum: 637b9adadf7acce5c70e5d327a725b13 Database to create: workshop Using curl from command line… curl -X POST -F email=yourEmail@server.com -F password= 637b9adadf7acce5c70e5d327a725b13 -F dbname=workshop https://my.metaio.com/REST/VisualSearch/addDatabase.php Server answers with… <?xml version=„1.0“ ?> <Response><AddedDatabase>workshop</AddedDatabase></Response>
    • Getting started with Visual Search 2. Add image: Using curl from command line… curl -X POST -F email=yourEmail@server.com -F password= 637b9adadf7acce5c70e5d327a725b13 -F dbname=workshop -F trackable=@metaioman.jpg (-F identifier=arbitraryName -F metadata=<validJSONObject>) https://my.metaio.com/REST/VisualSearch/addItem.php Server answers with… <?xml version=„1.0“ ?><Response><AddedItem Name=„metaioman.jpg“></AddedItema></Response>
    • Getting started with Visual Search 3. Allow an application to access this database: Using curl from command line… curl -X POST -F email=yourEmail@server.com -F password= 637b9adadf7acce5c70e5d327a725b13 -F dbname=workshop -F appIdentifier=com.metaio.insideARDemo https://my.metaio.com/REST/VisualSearch/addApplication.php Server answers with… <?xml version=„1.0“ ?><Response> <AddedApplication AppName=„com.metaio.insideARDemo“ DbName=„insideARDemo“ /> </Response>
    • Getting started with Visual Search 4. Build simple client application: • use the metaioSDK Example project and just change the application identifier • Register the application identifier in my.metaio.com „My Apps“ section • Make simple App logic in AREL/ javascript (logic.js): // start Visual Search once arel.sceneReady(function() { arel.Scene.requestVisualSearch(„workshop",…); arel.Events.setListener(arel.Scene, myCallback); }); // print the result function myCallback(type, result) { if (type == arel.Events.ONVISUALSEARCHRESULT) { for (int i = 0; i < result.size; i++) { console.log(result[i].identifier); } } }
    • Image constraints Which images to put into the database? Constraints are in general the same as for image trackable for client-only applications: http://dev.metaio.com/sdk/tracking-config/create-image-trackable/ Image size: 400 x 400 200 x 200 OK … OK
    • Image constraints Important for good tracking quality: http://dev.metaio.com/sdk/tracking-config/create-image-trackable/ Select a region which is a little bit smaller than the pre-print version, then upload this to your Visual Search Database. (Required for good tracking on client, not for good detection on server)
    • Q&A Q&A & wishes & more ... Additional resources and information: http://dev.metaio.com/junaio/continuous-visualsearch/general-information/
    • www.metaio.com