An Improved Small Target Detection Algorithm Based on YOLOv8s

Due to challenges such as the small size of targets, complex backgrounds, limited feature extraction capa-bilities, and frequent false positives and false negatives, traditional detection algorithms often perform poorly in small object detection tasks. To address these challenges, this pa¬per propos...

Full description

Saved in:
Bibliographic Details
Main Authors: G. Ma, C. Xu, Z. Xu, X. Song
Format: Article
Language:English
Published: Spolecnost pro radioelektronicke inzenyrstvi 2025-06-01
Series:Radioengineering
Subjects:
Online Access:https://www.radioeng.cz/fulltexts/2025/25_02_0206_0223.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850208424035352576
author G. Ma
C. Xu
Z. Xu
X. Song
author_facet G. Ma
C. Xu
Z. Xu
X. Song
author_sort G. Ma
collection DOAJ
description Due to challenges such as the small size of targets, complex backgrounds, limited feature extraction capa-bilities, and frequent false positives and false negatives, traditional detection algorithms often perform poorly in small object detection tasks. To address these challenges, this pa¬per proposes an enhanced small object detection algorithm, SOD-YOLO, based on YOLOv8s. First, the S_C2f_CAFM module is integrated into the feature extraction network, enabling the effective capture of fine-grained local features and broad contextual information, while simultaneously reducing model parameters and computational complexity. Second, in the feature fusion stage, the redesigned bidirectional feature pyramid network employs a spatial context awareness module to extract key features, adding a top-down path to optimize feature fusion and enhance discriminative information. In the Neck section, the D_C2f_MSPA module is introduced, which, while being lightweight, accurately models channel dependencies in feature maps, effectively reducing both false positives and false negatives for small objects. Finally, the inclusion of Normalized Wasserstein Distance (NWD) further improves detection accuracy and reduces the model’s sensitivity to small positional deviations in small objects. Experimental results on the DOTAv1.0, VisDrone2019, and TT100K datasets confirm that SOD-YOLO achieves excellent performance, demonstrating the effectiveness of the modifications made to the original YOLOv8 model.
format Article
id doaj-art-befac3ef52c14fbd8750e3ea0f9aa0d2
institution OA Journals
issn 1210-2512
language English
publishDate 2025-06-01
publisher Spolecnost pro radioelektronicke inzenyrstvi
record_format Article
series Radioengineering
spelling doaj-art-befac3ef52c14fbd8750e3ea0f9aa0d22025-08-20T02:10:14ZengSpolecnost pro radioelektronicke inzenyrstviRadioengineering1210-25122025-06-01342206223An Improved Small Target Detection Algorithm Based on YOLOv8sG. MaC. XuZ. XuX. SongDue to challenges such as the small size of targets, complex backgrounds, limited feature extraction capa-bilities, and frequent false positives and false negatives, traditional detection algorithms often perform poorly in small object detection tasks. To address these challenges, this pa¬per proposes an enhanced small object detection algorithm, SOD-YOLO, based on YOLOv8s. First, the S_C2f_CAFM module is integrated into the feature extraction network, enabling the effective capture of fine-grained local features and broad contextual information, while simultaneously reducing model parameters and computational complexity. Second, in the feature fusion stage, the redesigned bidirectional feature pyramid network employs a spatial context awareness module to extract key features, adding a top-down path to optimize feature fusion and enhance discriminative information. In the Neck section, the D_C2f_MSPA module is introduced, which, while being lightweight, accurately models channel dependencies in feature maps, effectively reducing both false positives and false negatives for small objects. Finally, the inclusion of Normalized Wasserstein Distance (NWD) further improves detection accuracy and reduces the model’s sensitivity to small positional deviations in small objects. Experimental results on the DOTAv1.0, VisDrone2019, and TT100K datasets confirm that SOD-YOLO achieves excellent performance, demonstrating the effectiveness of the modifications made to the original YOLOv8 model.https://www.radioeng.cz/fulltexts/2025/25_02_0206_0223.pdfyolov8small object detectionattention mechanismfeature fusionloss function
spellingShingle G. Ma
C. Xu
Z. Xu
X. Song
An Improved Small Target Detection Algorithm Based on YOLOv8s
Radioengineering
yolov8
small object detection
attention mechanism
feature fusion
loss function
title An Improved Small Target Detection Algorithm Based on YOLOv8s
title_full An Improved Small Target Detection Algorithm Based on YOLOv8s
title_fullStr An Improved Small Target Detection Algorithm Based on YOLOv8s
title_full_unstemmed An Improved Small Target Detection Algorithm Based on YOLOv8s
title_short An Improved Small Target Detection Algorithm Based on YOLOv8s
title_sort improved small target detection algorithm based on yolov8s
topic yolov8
small object detection
attention mechanism
feature fusion
loss function
url https://www.radioeng.cz/fulltexts/2025/25_02_0206_0223.pdf
work_keys_str_mv AT gma animprovedsmalltargetdetectionalgorithmbasedonyolov8s
AT cxu animprovedsmalltargetdetectionalgorithmbasedonyolov8s
AT zxu animprovedsmalltargetdetectionalgorithmbasedonyolov8s
AT xsong animprovedsmalltargetdetectionalgorithmbasedonyolov8s
AT gma improvedsmalltargetdetectionalgorithmbasedonyolov8s
AT cxu improvedsmalltargetdetectionalgorithmbasedonyolov8s
AT zxu improvedsmalltargetdetectionalgorithmbasedonyolov8s
AT xsong improvedsmalltargetdetectionalgorithmbasedonyolov8s