FEATURED WORK
LATEST NEWS

Diverse Ensembles and Random Projections for Robust Real-Time Visual Object Tracking


Tracking by detection techniques have been gaining popularity and showing promising results. They use samples classified in previous frames to detect an object in a new frame. We propose a novel real-time ensemble approach to tracking by detection where we create a diverse ensemble using random projections to select strong and diverse sets of compressed features. The proposed ensemble tracker significantly improves the accuracy of tracking while not using any additional information than that available to the single classifier. more info

Novel P-N Learning and Structural Constraints for Enhanced Target Tracking in UAV Imagery


We propose improved automatic moving target detection and tracking framework that is suitable for UAV imagery. The framework is comprised of motion compensation, target state estimation, and overlap-rate-based data association. Finally, P-N learning is used to maintain target appearance by utilizing novel structural constraints. The proposed framework enables to recapture targets after being out of camera field of view and helps discriminating between targets with similar appearance while alleviating drift problems. more info

Recent Grants:

  • Subsidies Mobile Wallet: Use of mobile phones and biometrics for secured subsidy distribution, Academy for Scientific Research and Technology in collaboration with Orange Labs Cairo, 2013-2015, EGP 3 million ($430,000).

  • TraffiSense-Pro: Development of smart vision-based traffic sensors,Academy for Scientific Research and Technology, 2013-2015, EGP 1.2 million ($175,000).

  • Recent Publications:

  • A. Karali, M. ElHelw “Motion History of Skeletal Volumes and Temporal change in Bounding Volume Fusion for Human Action Recognition”, in Multimodal Pattern Recognition of Social Signals in Human-Computer-Interaction, Springer Berlin Heidelberg, 2013.

  • M. Taie, K. Nasr, M. ElHelw, “ITS Navigation and Live Timetables for the Blind Based on RFID Robotic Localization Algorithms and ZigBee Broadcasting”,IEEE Robotics and Biomimetics (ROBIO), Shenzen, China, November 2013.

  • A. Salaheldin, S. Maher, M. ElHelw, “Robust Real-Time Tracking with Diverse Ensembles and Random Projections”,International Conference on Computer Vision (ICCV), Workshop on Visual Object Tracking (VOT), Sydney, Australia, December 2013.

  • M. Siam, M. ElHelw, “Enhanced Target Tracking in Aerial Imagery with P-N Learning and Structural Constraints”, International Conference on Computer Vision (ICCV), Workshop on Computer Vision in Vehicle Technology, Sydney, Australia, December 2013.


  •  

    Sensor Explorer: Monitoring and 3D Visualization for Ubiquitous Sensing Deployments

    Recent developments in wireless sensor networks have ushered in novel ubiquitous computing applications based on distributed large-scale data acquisition and interactive interpretation. Monitoring the state of ubiquitous sensing systems and visualization of collected information are challenging issues, however. This work presents a comprehensive monitoring and visualization framework for ubiquitous sensing deployments called Sensor Explorer. The framework provides an effective user interface for deploying and administering ubiquitous networks as well as for enabling interactive visualization and knowledge discovery in sensing data. more info

     

     

    Real-Time Vehicles Detection and Tracking with Haar-Like Features and Compressive Sensing

    This work presents real-time vision-based framework that detects and tracks vehicles from stationary camera with certain region of interest (ROI). The framework consists of vehicle detection, tracking, and data association steps. Vehicles are
    detected using Haar-like features while appearance-based compressive sensing model is built to keep a track of detected vehicles. The framework uses detections to enhance tracking and vice versa to sustain excellent performance under a variety of challenges. more info

     

     

    Use of Body and Visual Sensor Networks for Pervasive Healthcare Monitoring and Improved Healthcare Delivery

    By integrating the strengths of ambient and wearable sensing, it is possible to provide true pervasive systems that can be used to accurately infer patient conditions based on activity and physiological parameters. This project investigates an effective solution for monitoring the wellbeing of the elderly, quantifying post-operative patient recovery and monitoring the progression of neurodegenerative diseases such as Parkinson’s. more info

    Copyright © 2010-2014 Ubiquitous & Visual Computing Group