Design of real-time object classification acceleration platform based on SDSoC

Bo Zhang; Yixin Liu; Yuxuan Zhang

doi:10.1117/12.3009413

11 October 2023 Design of real-time object classification acceleration platform based on SDSoC

Bo Zhang, Yixin Liu, Yuxuan Zhang

Proceedings Volume 12918, Fourth International Conference on Computer Science and Communication Technology (ICCSCT 2023); 129181Z (2023) https://doi.org/10.1117/12.3009413
Event: International Conference on Computer Science and Communication Technology (ICCSCT 2023), 2023, Wuhan, China

Abstract

This study introduces a hardware acceleration design scheme for real-time object classification using SDSoC development software, and compares the performance advantages and disadvantages of different optimization strategies. Unlike the traditional RTL design and HLS design, SDSoC adopts the whole system optimization compiler, which can provide automatic data connection and other functions, so that developers can invest more energy in algorithm optimization, reduce the development difficulty and speed up the development time. This study uses a convolutional neural network to classify objects and optimizes the network according to hardware performance. A real-time video acquisition and display system was built in Vivado. This study builds a complete system in SDSoC, and focuses on optimizing convolution operation and data transmission path according to the running time of each function in the builtin TCF profiler. The effects of different acceleration function settings and network quantization on performance and resource utilization were studied. Finally, different optimization strategies are compared and evaluated, and the best design scheme achieves 3.599 W power consumption and 9.24 GOP/s per watt performance.

(2023) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Bo Zhang, Yixin Liu, and Yuxuan Zhang "Design of real-time object classification acceleration platform based on SDSoC", Proc. SPIE 12918, Fourth International Conference on Computer Science and Communication Technology (ICCSCT 2023), 129181Z (11 October 2023); https://doi.org/10.1117/12.3009413

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available