DMET 1071 Seminar on Edge Computing for Computer Vision

Course Information

Abstract

  • Deep learning is currently widely used in computer vision and multimedia processing. End devices, such as single-board-computer (SBC), smartphones and Internet-of-Things sensors, are generating/ consuming huge amount of data. They also provide benefits in terms of privacy, bandwidth efficiency, and scalability. However, training of deep learning models needs advanced computation resources to run effectively. Enabling edge devices to run deep learning means to analyze and process the data in the same place where it has been generated. Developing edge computing models for computer vision applications is a challenging objective however it worthy to investigate the recent trends and future opportunities. In this seminar, we aim to address reviews of the current state of the art at the intersection of computer vision and edge computing. We will have insight on applications where deep learning is used at the network edge, discuss various approaches for quickly executing deep learning inference across a combination of end devices, edge servers, and the cloud, and describe the methods for training deep learning models across multiple edge devices. We will also discuss open challenges in terms of performance, benchmarks, and privacy.
RenewSession