Paper
21 March 2023 YOLOv5s-L: object detection for service robots
Yi Liu, Yumo Wang
Author Affiliations +
Proceedings Volume 12609, International Conference on Computer Application and Information Security (ICCAIS 2022); 126090H (2023) https://doi.org/10.1117/12.2671857
Event: International Conference on Computer Application and Information Security (ICCAIS 2022), 2022, ONLINE, ONLINE
Abstract
Object detection, as one of the key technologies of service robots, is becoming increasingly important. For this reason, this paper proposes an improved YOLOv5s-L algorithm based on YOLOv5s for service scenarios such as hotels and shopping malls. With the introduction of Ghost series modules, light weight is realized by reducing the number of parameters and calculations of the network. And different compression ratios for Ghostconv modules are compared. To optimize the algorithm, a large-field contextual feature integration module and the coordinate attention mechanism are introduced, which help to enhance the ability to gain information about small objects in the scene and increase the sensitivity to information such as the position and direction of the target. A qualitative analysis was carried out for the Ghostconv modules with different compression ratios, and it was concluded that there is a large amount of redundancy in the deep features of YOLOv5s. The algorithm’s effectiveness was verified using two data sets, PASCAL VOC2007 and PASCAL VOC2012, and experiments indicate that our improved algorithm outperforms other compared methods.
© (2023) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Yi Liu and Yumo Wang "YOLOv5s-L: object detection for service robots", Proc. SPIE 12609, International Conference on Computer Application and Information Security (ICCAIS 2022), 126090H (21 March 2023); https://doi.org/10.1117/12.2671857
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Object detection

Detection and tracking algorithms

Robots

Convolution

Target detection

Feature extraction

Education and training

Back to Top