There are an increasing number of video analytics applications, such as metaverse games, auto-driving, etc. As a way to support these applications, much attention is being paid to edge-cloud video analytics systems. The standard pipeline is one in which the edge captures video contents; conducts preprocessing and sends intermediate results to the cloud to complete the analytics tasks. There are diverse requirements from applications, such as latency, privacy, etc., as well as diverse constraints from the resources in computing and communication. These make edge-cloud video analytics system designs challenging. In this talk, we will discuss issues related to edge-cloud video analytics system designs. In particular, we study edge-side acceleration using new hardware and the designs on resource-efficient privacy-preserving systems. We also present experiences in broadening the impact of research results into industry projects.

Co-sponsored by: New Jersey Institute of Technology

Speaker(s): Dr. Dan Wang,

Agenda:
There are an increasing number of video analytics applications, such as metaverse games, auto-driving, etc. As a way to support these applications, much attention is being paid to edge-cloud video analytics systems. The standard pipeline is one in which the edge captures video contents; conducts preprocessing and sends intermediate results to the cloud to complete the analytics tasks. There are diverse requirements from applications, such as latency, privacy, etc., as well as diverse constraints from the resources in computing and communication. These make edge-cloud video analytics system designs challenging. In this talk, we will discuss issues related to edge-cloud video analytics system designs. In particular, we study edge-side acceleration using new hardware and the designs on resource-efficient privacy-preserving systems. We also present experiences in broadening the impact of research results into industry projects.

Room: GITC 1400, Bldg: GITC 1400, 218 Central Avenue Newark, NJIT, Newark, New Jersey, United States, 07102, Virtual: https://events.vtools.ieee.org/m/330949