Online demo

Online demo is available on this page where you can evaluate accuracy of face verification procedure.

Introduction

StarTrinity Face Recognition Library is designed for biometric face login in web-based solutions. The library

  • Allows you to use human face as a password to authorize at your web site
  • Can be integrated into any ASP.NET or PHP web application
  • Is based on Silverlight 4 runtime, detects and verifies face in real time
  • Supports DirectShow-compatible web cameras.
  • Offers the API to detect face, build pattern of face and match face with pattern
  • Uses cloud server to build and store facial patterns
  • Accuracy is up to 99% since version 1.2 (FAR 0%, TRR 99%)

Integration into Silverlight 4 web application

First you need to register your application and get application Id and password. The credentials will be used to generate a security token. The security token has to generated on the server side using application Id and password. It is used to identify your application on the cloud server. Details of implementation can be found in the sample project.

 The library is compiled into StarTrinity.FaceRecognition.SL.dll Silverlight 4 binary file. It has to be added as a reference to your application and initialized with the security token. MVVM design pattern is used for this library. Each single view model represents a separate logic layer.

  • Web Camera View Model contains logic for capturing video and face detection. There are settings for resolution of video stream. Default resolution is 640x480.
  • Registration View Model takes snapshots for facial pattern. It uploads the snapshots to the cloud server.
  • Login View Model verifies face and provides status of verification. The status can be used in your application.

Last edited May 28, 2012 at 8:14 PM by asv128, version 8