Multi-resolution Rescored ByteTrack for Video Object Detection on Ultra-low-power Embedded Systems


METADATA ONLY
Loading...

Date

2024

Publication Type

Conference Paper

ETH Bibliography

yes

Citations

Altmetric
METADATA ONLY

Data

Rights / License

Abstract

This paper introduces Multi-Resolution Rescored ByteTrack (MR2-ByteTrack), a novel video object detection framework for ultra-low-power embedded processors. This method reduces the average compute load of an off-the-shelf Deep Neural Network (DNN) based object detector by up to 2.25x by alternating the processing of high-resolution images (320x320 pixels) with multiple down-sized frames (192x192 pixels). To tackle the accuracy degradation due to the reduced image input size, MR2-ByteTrack correlates the output detections over time using the ByteTrack tracker and corrects potential misclassification using a novel probabilistic Rescore algorithm. By interleaving two down-sized images for every high-resolution one as the input of different state-of-the-art DNN object detectors with our MR2-ByteTrack, we demonstrate an average accuracy increase of 2.16% and a latency reduction of 43% on the GAP9 microcontroller compared to a baseline frame-by-frame inference scheme using exclusively full-resolution images. Code available at: https://github.com/Bomps4/ Multi_Resolution_Rescored_ByteTrack

Publication status

published

Editor

Book title

2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)

Journal / series

Volume

Pages / Article No.

2182 - 2190

Publisher

IEEE

Event

2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW 2024)

Edition / version

Methods

Software

Geographic location

Date collected

Date created

Subject

Video object detection; Neural networks; Ultra-low-power; Embedded systems

Organisational unit

03996 - Benini, Luca / Benini, Luca check_circle

Notes

Funding

207913 - TinyTrainer: On-chip Training for TinyML devices (SNF)

Related publications and datasets