Speaker: Murad Qasaimeh, ECpE Graduate Student
Advisor: Phillip Jones
Title: Techniques for Efficient Processing of Computer Vision Algorithms on Reconfigurable Computing Architectures
Abstract: Computer vision algorithms empowered with the recent advances in deep learning play a fundamental role in solving many problems that seemed impossible just a decade ago. However, the computational complexity and memory footprint of these algorithms keep increasing to enhance accuracy and solve more complex problems. This put heavy load on computing platforms used to run these algorithms. Moreover, the exponential growth in computing platforms’ capability starts slowing down because of the saturation in Moore’s law, which makes the problem even more challenging. In this work, we focus on algorithmic and hardware optimization techniques for efficient processing of computer vision algorithms on reconfigurable architectures. The proposed techniques aim to improve architecture’s efficiency in terms of run-time, energy consumption, resource utilization and programmability.