Title: Detection and recognition of multiple QR codes based on YOLO_CBAM algorithm

Authors: Juntao Li; Meijuan Zhao; Zhenbo Qin; Ruiping Yuan; Anqiang Huang; Mengtao Li

Addresses: School of Information, Beijing Wuzi University, Beijing 101149, China; Beijing Key Laboratory of Intelligent Logistics Systems, Beijing 101149, China ' School of Information, Beijing Wuzi University, Beijing 101149, China; Beijing Key Laboratory of Intelligent Logistics Systems, Beijing 101149, China ' School of Information, Beijing Wuzi University, Beijing 101149, China; Beijing Key Laboratory of Intelligent Logistics Systems, Beijing 101149, China ' School of Information, Beijing Wuzi University, Beijing 101149, China; Beijing Key Laboratory of Intelligent Logistics Systems, Beijing 101149, China ' School of Economics and Management, Beijing Jiaotong University, Beijing 100044, China ' School of Business Administration, Northeast University of Finance and Economics, Dalian 116025, China

Abstract: Different forms of two-dimensional codes are already widely used in all aspects of people's production and life, and the demand for rapid detection and recognition technology of multiple QR codes in various complex scenarios is also increasing. The traditional QR code detection algorithm has a high miss detection rate. In this paper, we collected a little sample data of multiple QR codes and annotated them. We used the YOLOv3 and YOLOv5 algorithms to implement the detection of multiple QR codes. Then, we added CBAM to the YOLO algorithm, and added an angle prediction mechanism to improve their decoding and recognition effects. The experimental results showed that the YOLOv3_CBAM algorithm and the YOLOv5_CBAM algorithm improved by 0.63% and 8.89% about mAP@.5 and mAP@.5:.95 respectively for the multiple QR codes dataset, with a detection speed of 70 FPS and achieved real-time performance.

Keywords: detection and recognition of multiple QR codes; angle prediction; YOLOv3 algorithm; YOLOv5 algorithm; attention mechanism.

DOI: 10.1504/IJBIC.2024.137915

International Journal of Bio-Inspired Computation, 2024 Vol.23 No.3, pp.179 - 188

Received: 11 Nov 2022
Accepted: 02 Aug 2023
Published online: 08 Apr 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article