A Survey on Efficient Cloud Vision Analytics

PhD Qualifying Examination


Title: "A Survey on Efficient Cloud Vision Analytics"

by

Mr. Yiding WANG


Abstract:

Cloud computing has been empowering many emerging internet applications and 
shaping the software service in recent decades. Running applications in the 
datacenters and delivering them as services over the internet let developers 
efficiently handle the heterogeneity of various client devices and let users 
take advantage of the powerful cloud infrastructure. With the rapidly 
developing deep learning techniques, computer vision analytics have been 
achieving superior performance and are widely adopted in real-world 
applications, such as traffic monitoring and autonomous driving.

The large-scale deployments of cameras are ubiquitous today and generate 
enormous video data. Running complex vision analytics applications in the cloud 
datacenters is a common industry practice, and a growing number of systems are 
designed for this. However, the shared network infrastructure could be 
overloaded, especially for wireless and cellular networks.

This survey will first look into different vision analytics tasks and their 
status in cloud computing. Then we will investigate several recent system and 
networking research works on two aspects: how to efficiently run vision 
analytics in the datacenters and how to handle the edge-to-cloud barriers 
especially the bandwidth constraint in cloud vision analytics tasks.


Date:			Wednesday, 12 June 2019

Time:                  	10:15am - 12:15pm

Venue:                  Room 3494
                         Lifts 25/26

Committee Members:	Dr. Kai Chen (Supervisor)
 			Dr. Qifeng Chen (Chairperson)
 			Dr. Yangqiu Song
 			Dr. Wei Wang


**** ALL are Welcome ****