Publication Type

Conference Proceeding Article

Version

publishedVersion

Publication Date

1-2024

Abstract

We demonstrate criticality-aware canvas-based processing of multiple concurrent camera streams at the resource constrained edge to show substantial improvement in the accuracy-throughput trade-off. The proposed system focuses the available computation resources on select Regions of Interest (RoI) across all the camera streams by (i) extracting RoI from the input camera stream (ii) 2D bin packing the RoI on a canvas frame and (iii) batching and inferring upon these constructed composite canvas frames with a YOLOv5 object detection model. Our experiments show that such canvas-based processing can (i) sustain real-time processing throughput of 23 FPS per camera across 6 concurrent input camera streams (cumulatively 138 FPS) on a single NVIDIA Jetson TX2 representing a 475% increase in throughput, with (ii) negligible loss in accuracy as compared to a First Come First Serve (FCFS) baseline running full frame detections on the input camera streams.

Keywords

Canvas-based Processing, Edge Computation, Multi-Camera Systems

Discipline

Graphics and Human Computer Interfaces | Software Engineering

Research Areas

Data Science and Engineering

Areas of Excellence

Digital transformation

Publication

Proceedings of the 2024 16th International Conference on COMmunication Systems & NETworkS (COMSNETS), Bengaluru, India January 3-7

First Page

297

Last Page

299

ISBN

9798350383119

Identifier

10.1109/COMSNETS59351.2024.10427123

Publisher

IEEE

City or Country

Pistacataway, NJ

Embargo Period

8-26-2024

Copyright Owner and License

Authors

Additional URL

https://doi.org/10.1109/COMSNETS59351.2024.10427123

Download

Share

COinS