Daniel Palomino and Guilherme Correa
ViTech Research Group, Universidade Federal de Pelotas, Brazil
Enabling visual signal coding systems: from algorithm to hardware solutions
Monday, October 19th, 01:30 PM — 03:00 PM (GMT-3)
This presentation will be given in Portuguese.
Abstract
Visual signal coding has been improved in the past years by the development of new tools and algorithms together with the advances in the devices processing capabilities. Visual signal technologies such as ultra-high definition (UHD) video, 3D and omnidirectional 360-degree imaging, light fields and point clouds are now possible due to these advances. However, the computational complexity of visual signal coding solutions has increased dramatically, making the feasibility of such systems, especially those for real-time operation, a huge challenge. This talk will present challenges and solutions for enabling visual signal coding systems considering these recent technologies targeting algorithm-level and system-level solutions. The main challenge from the algorithm-level perspective is to reduce the computational complexity of the coding process without compromising coding efficiency. From the system-level perspective, the main goal is to provide low-power and high-throughput coding systems. The solutions presented in this talk include the use of heuristics, meta-heuristics and machine learning-based techniques for complexity reduction, and the design of hardware accelerators, memory hierarchies, system architectures and parallel computation for low-power and high-throughput visual signal coding.
Biography
Daniel Palomino is a professor at Center for Engineering (CENG), Federal University of Pelotas (UFPel), Brazil, since December 2014. He is also a researcher with the Video Technology Research Group (ViTech) and the Group of Architecture and Integrated Circuits (GACI). He received his Masters (February 2013) and Ph.D. (April 2017) in Computer Science from the Federal University of Rio Grande do Sul (UFRGS). His research interests are computer architectures, power/thermal/energy-efficient systems, FPGAs and embedded systems. The main focus of his research is on exploiting solutions at different computer layers (hardware, software and system) to enable coding systems for new video technologies such as UHD, omnidirectional 360-degree and light fields.
Guilherme Corrêa is a professor at the Technology Development Center (CDTec) from the Federal University of Pelotas (UFPel), Brazil. He is also a researcher with the Video Technology Research Group (ViTech) and the Group of Architectures and Integrated Circuits (GACI). Prof. Corrêa received the Ph.D. degree in Electrical and Computer Engineering from the University of Coimbra, Portugal (2015), the M.Sc. degree in Computer Science from the Federal University of Rio Grande do Sul (UFRGS) Brazil (2010), and the B.Sc. degree in Computer Science from the Federal University of Pelotas (UFPel), Brazil (2008). His research interests include visual signal compression algorithms, complexity reduction for video coding/transcoding based on heuristics and machine learning techniques, and digital systems design for multimedia technologies.