HIGH DENSITY DEEP LEARNING-LITE WITH INTEL OPENVINO™

YEW SHUN, OOI (2018) HIGH DENSITY DEEP LEARNING-LITE WITH INTEL OPENVINO™. [Final Year Project]

[thumbnail of Final Dissertation_19812.pdf] PDF
Final Dissertation_19812.pdf
Restricted to Registered users only

Download (1MB)

Abstract

Video is playing the vital role in Internet of Things (IoT) industry, as it is able to provide multi-dimensional data and handling complex task compared to typical sensor. Through image processing and video analytic, many industries such as factories, cities and offices will have improved in terms of the output’s efficiency and productivity. This technology can be further improved through implementing artificial intelligence, to yield faster and more accurate output. However, one of the biggest challenge of adapting video in various field is the hardware limitation. A typical video enabled IoT system usually require high performance specification, and with the addition of artificial intelligence algorithm, the requirement for the hardware performance is even higher. Intel had come out with a new product, Movidius Vision Processing Unit, which is able to run the video processing and machine learning with low power consumption while providing a high processing capability. This paper proposes a new add in card design powered with the VPU system on chip which is more suitable to adopt by other existing system, to improve the performance and reduce the system’s processor workload.

Item Type: Final Year Project
Departments / MOR / COE: Engineering > Electrical and Electronic
Depositing User: Mr Ahmad Suhairi Mohamed Lazim
Date Deposited: 20 Jun 2019 08:44
Last Modified: 20 Jun 2019 08:44
URI: http://utpedia.utp.edu.my/id/eprint/19199

Actions (login required)

View Item
View Item