Vision Processing Lab: Embedded & Real-time Systems for Machine Learning, Computer Vision & AI

VISpro unites professionals and researchers from the domains of Vision, Imaging & Signal Processing to develop innovative and practical systems aimed at solving the latest challenges being faced in these areas.

VISpro idealizes itself as a "system design" group aimed at bringing innovation and cutting edge research to product development in the defined areas.

The group is actively focused on FPGA based implementation of complex algorithms related to the aforementioned domains. Currently VISPRO is actively pursuing projects with local and international companies with projects related to Video Registration, Reconfigurable and real-time system development for immersive multimedia applications and multi-projector based Dome Projections for entertainment and military simulations.

News & Updates

VISpro wins Intel Heterogeneous Architecture Research Platform 2 (HARP-v2) award.

VISpro alumni Muhammad Abdullah Hanif wins the SIGDA Richard Newton Young Fellow Award.

Paper accepted at ISVLSI 2017. Title "Adaptive and Energy-Efficient Architectures for Machine Learning: Challenges, Opportunities, and Research Roadmap"

Postgraduate Fellowships Available @ VISpro Lab. Apply till 5th July 2017!

Papers Accepted in Tier-1 conferences: Design Automation Conference (DAC) 2017, and Design, Automation and Test in Europe Conference (DATE) 2017. For details, visit the publications page

Our Focus

Approximate Computing

Applications of approximate computing in the field of computer vision and image processing

Vision Based Systems

Developing algorithms for different vision applications such as Super Resolution, ADAS, Autonomous Driving Cars, and Facial Recognition

Energy Efficient Accelerators for Machine Learning

Adaptive and Energy-Efficient Architectures for Machine Learning, specifically for Convolutional Neural Networks, and Deep Neural Networks

Reconfigurable Accelerators

HW/SW Co-Design, FPGA based design for Machine Learning, and Computer Vision