Q-YOLO: Efficient Inference for Real-Time Object Detection

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Real-time object detection plays a vital role in various computer vision applications. However, deploying real-time object detectors on resource-constrained platforms poses challenges due to high computational and memory requirements. This paper describes a low-bit quantization method to build a highly efficient one-stage detector, dubbed as Q-YOLO, which can effectively address the performance degradation problem caused by activation distribution imbalance in traditional quantized YOLO models. Q-YOLO introduces a fully end-to-end Post-Training Quantization (PTQ) pipeline with a well-designed Unilateral Histogram-based (UH) activation quantization scheme, which determines the maximum truncation values through histogram analysis by minimizing the Mean Squared Error (MSE) quantization errors. Extensive experiments on the COCO dataset demonstrate the effectiveness of Q-YOLO, outperforming other PTQ methods while achieving a more favorable balance between accuracy and computational cost. This research contributes to advancing the efficient deployment of object detection models on resource-limited edge devices, enabling real-time detection with reduced computational and memory overhead.

Original languageEnglish
Title of host publicationPattern Recognition - 7th Asian Conference, ACPR 2023, Proceedings
EditorsHuimin Lu, Michael Blumenstein, Sung-Bae Cho, Cheng-Lin Liu, Yasushi Yagi, Tohru Kamiya
PublisherSpringer Science and Business Media Deutschland GmbH
Pages307-321
Number of pages15
ISBN (Print)9783031476648
DOIs
StatePublished - 2023
Event7th Asian Conference on Pattern Recognition, ACPR 2023 - Kitakyushu, Japan
Duration: 5 Nov 20238 Nov 2023

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume14408 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference7th Asian Conference on Pattern Recognition, ACPR 2023
Country/TerritoryJapan
CityKitakyushu
Period5/11/238/11/23

Keywords

  • Post-training Quantization
  • Real-time Object Detection

Fingerprint

Dive into the research topics of 'Q-YOLO: Efficient Inference for Real-Time Object Detection'. Together they form a unique fingerprint.

Cite this