iPronto Systems

Creative ideas, Spirited Team

Projects

  • Voice Over IP - Interactive Voice Response (IVR)
  • Voice Over IP - Creating Media Server Platform
  • Video/Audio MIX
  • Vehicle Velocity Detection

Voice Over IP


Interactive Voice Response (IVR):

The project was designing IVR to use Host Media Processing instead of on-chip processors for serving VXML based IVR(s). Project's result have following features:

  • Uses SIP as Signalling protocol
  • Uses VXML as programming language for IVR applications
  • Uses Host Media Processing instead of Media Boards (such as Dialogin Cards)
  • Enterprise, high-load support to use in real-world communication environments
  • Database support for logging user accesses
  • Extra Telnet server for configuration and manipulation
  • Extra logging features through syslog
  • TTS and ASR engine loaders
  • Support VXML plugins

Creating Media Server Platform

The project aim was extending first project result to support all media server features through application. By extending platform the first project support extra features:

  • Support fax and conferencing
  • Enterprise message-server for better support
  • Automatic reference counting
  • Using COM-like Interface-Based model

Video/Audio MIX

The project was developped by a request from StreamUK company. They wanted an application which could multiplex four AVI files in 2x2 view. Additionally, the application should create audio track which contains all audio tracks of source video. We used DirectShow internal features for decompressing video and compressing it. For image processing, video multiplexing and logo (or banner) addition we used OpenCV library. The application contains web based control which queries current active jobs, their progress and ability to stop them. This project is part of following website which is done to promote a Sony product: http://www.walkmanproject.com/


Vehicle Velocity Detection

I had done this project as my MSc project in University. We started that because of a problem of Iranian Police Force in detecting speed of cars in highways and roads using old Sonar and Laser sensors. The problem that, sometimes the sensor mistakes and assumes front weels of a car (say Car 1) and rear weels of another car (say Car 2) as a Car. So, it reports an invalid speed. We experienced following features in development of this project:

  • Object Tracking and Extraction
  • Different Background Modeling techniques, and presenting new techniques
  • Real world mapping and extracting real position of objects using Homography and Inverse Prespective Mapping
  • Extracting Velocity of objects in scene
  • Recording result of algorithm in AVI files using libquicktime
  • Using SDL for presenting result of algoithm in graphical window (Realtime)

By presenting techniques used in this project, we published three papers in different conferences:

  • R Zabihollahi, M Soryani, "Vehicle Velocity Detection System Based on Real-Time Motion Tracking," Accepted by Machine Vision and Image processing, 2007.
  • R Zabihollahi, M Soryani, "Vehicle Shadow Exclusion for a Vehicle Velocity Detection System," Submitted to International Multi-Conference of Engineers and Computer Scientists, 2007.
  • R Zabihollahi, M Soryani, "Edge based tracking," Submitted to Iranian Conference on Electrical Engineering, 2007.