High-Performance Computing for Computer Vision

One Day Meeting: High-Performance Computing for Computer Vision

Wednesday 22 May 2019

Chairs: Nicoletta Noceti & Giuseppe Ciaccio

Keynote Speakers

Meeting Report

After the meeting the organisers prepare a short summary of the meeting. The report is available here.

In memory of Giuseppe Ciaccio, open mind, brilliant scientist.

Programme

The last decades have witnessed amazing progress in the field of Computer Vision, where state-of-art methodologies provide astonishing performance. At the root of this growth is the conjunction between advances in computer architecture and availability of large amounts of data, that have triggered new lines of research while shedding renewed light on classical Computer Vision tasks, considered intractable from a practical standpoint until a few years ago. New high-performance computing platforms and paradigms (many-core processors, FPGA accelerators, large-scale clusters, cloud computing) is gaining momentum and becoming mainstream. The interplay between algorithms/methods for Computer Vision and high-performance computing architecture/programming thus continues to provide interesting challenges to researchers in the field. The goal of this one-day meeting is to provide a view on trends and issues in the use of modern high-performance and large-scale computing platforms for the challenges of Computer Vision.

09:30: Keynote: Tae-Kiun Kim “Machine learning for 3D vision”.
10:10: Oscar Rahnama “Developing Real-Time Stereo Algorithms for FPGA-based Platforms”
10:30: Alessio Tonioni “Real-time self-adaptive deep stereo”
10:50: Coffee Break
11:20: Matteo Poggi “Real-time monocular depth estimation without GPU”
11:40: Armin Mustafa “4D Vision in the wild”
12:00: Keynote: Juan Gomez Luna “From GPGPUs to Processing-in-Memory: High Performance and Energy Efficiency for Computer Vision Workloads”
12:40: Lunch
13:50: keynote: Marta Betcke “High-resolution tomographic imaging with learned priors”
14:30: Olivia Wilson “Mega Maps: Large scale deep learning of aerial imagery”
14:50: Bolin Pan “Photoacoustic Imaging using Curvelet Sparsity with Optical Flow Constraints”
15:10: Coffee Break
15:40: Nargiza Djurabekova “Reconstructing ankle and foot dynamics using optical flow”.
16:00: Elena Nicora “ Why High-performance computing is important in robotics?”


Meeting Location

British Computer Society (BCS), 5 Southampton St, London WC2E 7HA