DSR-YOLO: A lightweight and efficient YOLOv8 model for enhanced pedestrian detection

This paper presents DSR-YOLO, a pedestrian detection network that addresses critical challenges, such as scale variations and complex backgrounds. Built on the lightweight YOLOv8n architecture, it incorporates DCNv4 modules to enhance the detection rates and reduce missed detections by effectively l...

Full description

Saved in:
Bibliographic Details
Main Authors: Mustapha Oussouaddi, Omar Bouazizi, Aimad El mourabit, Zine el Abidine Alaoui Ismaili, Yassine Attaoui, Mohamed Chentouf
Format: Article
Language:English
Published: KeAi Communications Co. Ltd. 2025-01-01
Series:Cognitive Robotics
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2667241325000096
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849700493416202240
author Mustapha Oussouaddi
Omar Bouazizi
Aimad El mourabit
Zine el Abidine Alaoui Ismaili
Yassine Attaoui
Mohamed Chentouf
author_facet Mustapha Oussouaddi
Omar Bouazizi
Aimad El mourabit
Zine el Abidine Alaoui Ismaili
Yassine Attaoui
Mohamed Chentouf
author_sort Mustapha Oussouaddi
collection DOAJ
description This paper presents DSR-YOLO, a pedestrian detection network that addresses critical challenges, such as scale variations and complex backgrounds. Built on the lightweight YOLOv8n architecture, it incorporates DCNv4 modules to enhance the detection rates and reduce missed detections by effectively learning key pedestrian features. A new head component enables detection across various scales, whereas RFB modules improve accuracy for smaller or occluded objects. Additionally, we enhance the initial C2f layers with a modified block that integrates SimAM and DCNv4, minimizing the background noise and sharpening the focus on the relevant features. A second version of the C2f block using SimAM and standard convolutions ensures robust feature extraction in deeper layers with optimized computational efficiency. The WIoUv3 loss function was utilized to reduce the regression loss associated with bounding boxes, further boosting the performance. Evaluated on the CityPersons dataset, DSR-YOLO outperformed YOLOv8n with a 14.9 % increase in mAP@50 and 6.3 % increase in mAP@50:95, while maintaining competitive FLOPS, parameter counts, and inference speed.
format Article
id doaj-art-37d1ab422b1241b5b25990fb41f3c4aa
institution DOAJ
issn 2667-2413
language English
publishDate 2025-01-01
publisher KeAi Communications Co. Ltd.
record_format Article
series Cognitive Robotics
spelling doaj-art-37d1ab422b1241b5b25990fb41f3c4aa2025-08-20T03:18:15ZengKeAi Communications Co. Ltd.Cognitive Robotics2667-24132025-01-01515216510.1016/j.cogr.2025.04.001DSR-YOLO: A lightweight and efficient YOLOv8 model for enhanced pedestrian detectionMustapha Oussouaddi0Omar Bouazizi1Aimad El mourabit2Zine el Abidine Alaoui Ismaili3Yassine Attaoui4Mohamed Chentouf5Corresponding author.; SDE Team, ENSA of Tangier, Abdelmalek Essaâdi University, Tangier, MoroccoSDE Team, ENSA of Tangier, Abdelmalek Essaâdi University, Tangier, MoroccoSDE Team, ENSA of Tangier, Abdelmalek Essaâdi University, Tangier, MoroccoICES Team, Mohammed V University, Rabat, MoroccoSDE Team, ENSA of Tangier, Abdelmalek Essaâdi University, Tangier, Morocco; Siemens EDA/CSD Calypto - Synthesis Solutions, Rabat, MoroccoSiemens EDA/CSD Calypto - Synthesis Solutions, Rabat, MoroccoThis paper presents DSR-YOLO, a pedestrian detection network that addresses critical challenges, such as scale variations and complex backgrounds. Built on the lightweight YOLOv8n architecture, it incorporates DCNv4 modules to enhance the detection rates and reduce missed detections by effectively learning key pedestrian features. A new head component enables detection across various scales, whereas RFB modules improve accuracy for smaller or occluded objects. Additionally, we enhance the initial C2f layers with a modified block that integrates SimAM and DCNv4, minimizing the background noise and sharpening the focus on the relevant features. A second version of the C2f block using SimAM and standard convolutions ensures robust feature extraction in deeper layers with optimized computational efficiency. The WIoUv3 loss function was utilized to reduce the regression loss associated with bounding boxes, further boosting the performance. Evaluated on the CityPersons dataset, DSR-YOLO outperformed YOLOv8n with a 14.9 % increase in mAP@50 and 6.3 % increase in mAP@50:95, while maintaining competitive FLOPS, parameter counts, and inference speed.http://www.sciencedirect.com/science/article/pii/S2667241325000096Pedestrian detectionDCNv4SimAMYOLOv8RFBWIoUv3
spellingShingle Mustapha Oussouaddi
Omar Bouazizi
Aimad El mourabit
Zine el Abidine Alaoui Ismaili
Yassine Attaoui
Mohamed Chentouf
DSR-YOLO: A lightweight and efficient YOLOv8 model for enhanced pedestrian detection
Cognitive Robotics
Pedestrian detection
DCNv4
SimAM
YOLOv8
RFB
WIoUv3
title DSR-YOLO: A lightweight and efficient YOLOv8 model for enhanced pedestrian detection
title_full DSR-YOLO: A lightweight and efficient YOLOv8 model for enhanced pedestrian detection
title_fullStr DSR-YOLO: A lightweight and efficient YOLOv8 model for enhanced pedestrian detection
title_full_unstemmed DSR-YOLO: A lightweight and efficient YOLOv8 model for enhanced pedestrian detection
title_short DSR-YOLO: A lightweight and efficient YOLOv8 model for enhanced pedestrian detection
title_sort dsr yolo a lightweight and efficient yolov8 model for enhanced pedestrian detection
topic Pedestrian detection
DCNv4
SimAM
YOLOv8
RFB
WIoUv3
url http://www.sciencedirect.com/science/article/pii/S2667241325000096
work_keys_str_mv AT mustaphaoussouaddi dsryoloalightweightandefficientyolov8modelforenhancedpedestriandetection
AT omarbouazizi dsryoloalightweightandefficientyolov8modelforenhancedpedestriandetection
AT aimadelmourabit dsryoloalightweightandefficientyolov8modelforenhancedpedestriandetection
AT zineelabidinealaouiismaili dsryoloalightweightandefficientyolov8modelforenhancedpedestriandetection
AT yassineattaoui dsryoloalightweightandefficientyolov8modelforenhancedpedestriandetection
AT mohamedchentouf dsryoloalightweightandefficientyolov8modelforenhancedpedestriandetection