Ablation experiments on custom dataset.

<div><p>Pyrotechnic detection has always been one of the critical issues in blasting safety. Due to the complex environment of blasting sites, irregular detonator wire postures, and the differences in object scales, making the detection of pyrotechnics more challenging. To address these...

Full description

Saved in:
Bibliographic Details
Main Author: LinNa Li (5665885) (author)
Other Authors: Han Gao (486886) (author), JunYi Lu (21162047) (author), XiaoXiao Xu (21162050) (author)
Published: 2025
Subjects:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1852021132827295744
author LinNa Li (5665885)
author2 Han Gao (486886)
JunYi Lu (21162047)
XiaoXiao Xu (21162050)
author2_role author
author
author
author_facet LinNa Li (5665885)
Han Gao (486886)
JunYi Lu (21162047)
XiaoXiao Xu (21162050)
author_role author
dc.creator.none.fl_str_mv LinNa Li (5665885)
Han Gao (486886)
JunYi Lu (21162047)
XiaoXiao Xu (21162050)
dc.date.none.fl_str_mv 2025-04-22T17:39:26Z
dc.identifier.none.fl_str_mv 10.1371/journal.pone.0318172.t007
dc.relation.none.fl_str_mv https://figshare.com/articles/dataset/Ablation_experiments_on_custom_dataset_/28842788
dc.rights.none.fl_str_mv CC BY 4.0
info:eu-repo/semantics/openAccess
dc.subject.none.fl_str_mv Biochemistry
Space Science
Environmental Sciences not elsewhere classified
Biological Sciences not elsewhere classified
Information Systems not elsewhere classified
parallel attention mechanism
novel convolutional block
new convolutional block
iou loss function
improved algorithm based
model &# 8217
improved model achieved
varying aspect ratios
experimental results show
depthwise separable convolution
yolov8 </ p
improving feature extraction
blasting sites based
integrate wavelet convolution
average precision increase
separable convolution
blasting sites
model uses
varying scales
results indicate
recall increase
blasting safety
voc2012 dataset
scale convolutions
paper proposes
overlap calculations
object scales
neck network
irregular shapes
effective solution
detonator wires
detect objects
critical issues
complex environment
complex backgrounds
c2f module
built dataset
bounding boxes
dc.title.none.fl_str_mv Ablation experiments on custom dataset.
dc.type.none.fl_str_mv Dataset
info:eu-repo/semantics/publishedVersion
dataset
description <div><p>Pyrotechnic detection has always been one of the critical issues in blasting safety. Due to the complex environment of blasting sites, irregular detonator wire postures, and the differences in object scales, making the detection of pyrotechnics more challenging. To address these challenges, this paper proposes an improved algorithm based on a multi-scale parallel attention mechanism and wavelet-separable convolution, called WA-YOLO. First, we integrate wavelet convolution into depthwise separable convolution and propose a novel convolutional block (WSDConv, Wavelet Separable Depthwise Convolution). This new convolutional block is added to the model’s backbone, improving feature extraction while also lowering computational parameters. Furthermore, we introduce an improved Cross Stage Partial (CSP) structure by combining multi-scale convolutions with a parallel attention mechanism, embedding it into the C2f module of the neck network to improve the model’s ability to detect objects of varying scales in complex backgrounds. To tackle the detection accuracy drop caused by the irregular shapes and varying aspect ratios of detonator wires, the model uses the Wise-IoU loss function. This enhances the model’s generalization and robustness by improving the precision of overlap calculations for bounding boxes. The experimental results show that the improved model achieved an average precision increase of 12.6% on the self-built dataset, particularly with an average precision increase of 8.3% in the detection of detonators. Additionally, the model performance also improved on the VOC2012 dataset, with a recall increase of 1.3% and an average precision increase of 1.6%. These results indicate that the proposed model exhibits strong generalization capabilities, can work effectively across different datasets, and provides an effective solution to the challenges of target detection in blasting environments.</p></div>
eu_rights_str_mv openAccess
id Manara_ace77ed2fa4cdbd331ff94b2bd2e42e7
identifier_str_mv 10.1371/journal.pone.0318172.t007
network_acronym_str Manara
network_name_str ManaraRepo
oai_identifier_str oai:figshare.com:article/28842788
publishDate 2025
repository.mail.fl_str_mv
repository.name.fl_str_mv
repository_id_str
rights_invalid_str_mv CC BY 4.0
spelling Ablation experiments on custom dataset.LinNa Li (5665885)Han Gao (486886)JunYi Lu (21162047)XiaoXiao Xu (21162050)BiochemistrySpace ScienceEnvironmental Sciences not elsewhere classifiedBiological Sciences not elsewhere classifiedInformation Systems not elsewhere classifiedparallel attention mechanismnovel convolutional blocknew convolutional blockiou loss functionimproved algorithm basedmodel &# 8217improved model achievedvarying aspect ratiosexperimental results showdepthwise separable convolutionyolov8 </ pimproving feature extractionblasting sites basedintegrate wavelet convolutionaverage precision increaseseparable convolutionblasting sitesmodel usesvarying scalesresults indicaterecall increaseblasting safetyvoc2012 datasetscale convolutionspaper proposesoverlap calculationsobject scalesneck networkirregular shapeseffective solutiondetonator wiresdetect objectscritical issuescomplex environmentcomplex backgroundsc2f modulebuilt datasetbounding boxes<div><p>Pyrotechnic detection has always been one of the critical issues in blasting safety. Due to the complex environment of blasting sites, irregular detonator wire postures, and the differences in object scales, making the detection of pyrotechnics more challenging. To address these challenges, this paper proposes an improved algorithm based on a multi-scale parallel attention mechanism and wavelet-separable convolution, called WA-YOLO. First, we integrate wavelet convolution into depthwise separable convolution and propose a novel convolutional block (WSDConv, Wavelet Separable Depthwise Convolution). This new convolutional block is added to the model’s backbone, improving feature extraction while also lowering computational parameters. Furthermore, we introduce an improved Cross Stage Partial (CSP) structure by combining multi-scale convolutions with a parallel attention mechanism, embedding it into the C2f module of the neck network to improve the model’s ability to detect objects of varying scales in complex backgrounds. To tackle the detection accuracy drop caused by the irregular shapes and varying aspect ratios of detonator wires, the model uses the Wise-IoU loss function. This enhances the model’s generalization and robustness by improving the precision of overlap calculations for bounding boxes. The experimental results show that the improved model achieved an average precision increase of 12.6% on the self-built dataset, particularly with an average precision increase of 8.3% in the detection of detonators. Additionally, the model performance also improved on the VOC2012 dataset, with a recall increase of 1.3% and an average precision increase of 1.6%. These results indicate that the proposed model exhibits strong generalization capabilities, can work effectively across different datasets, and provides an effective solution to the challenges of target detection in blasting environments.</p></div>2025-04-22T17:39:26ZDatasetinfo:eu-repo/semantics/publishedVersiondataset10.1371/journal.pone.0318172.t007https://figshare.com/articles/dataset/Ablation_experiments_on_custom_dataset_/28842788CC BY 4.0info:eu-repo/semantics/openAccessoai:figshare.com:article/288427882025-04-22T17:39:26Z
spellingShingle Ablation experiments on custom dataset.
LinNa Li (5665885)
Biochemistry
Space Science
Environmental Sciences not elsewhere classified
Biological Sciences not elsewhere classified
Information Systems not elsewhere classified
parallel attention mechanism
novel convolutional block
new convolutional block
iou loss function
improved algorithm based
model &# 8217
improved model achieved
varying aspect ratios
experimental results show
depthwise separable convolution
yolov8 </ p
improving feature extraction
blasting sites based
integrate wavelet convolution
average precision increase
separable convolution
blasting sites
model uses
varying scales
results indicate
recall increase
blasting safety
voc2012 dataset
scale convolutions
paper proposes
overlap calculations
object scales
neck network
irregular shapes
effective solution
detonator wires
detect objects
critical issues
complex environment
complex backgrounds
c2f module
built dataset
bounding boxes
status_str publishedVersion
title Ablation experiments on custom dataset.
title_full Ablation experiments on custom dataset.
title_fullStr Ablation experiments on custom dataset.
title_full_unstemmed Ablation experiments on custom dataset.
title_short Ablation experiments on custom dataset.
title_sort Ablation experiments on custom dataset.
topic Biochemistry
Space Science
Environmental Sciences not elsewhere classified
Biological Sciences not elsewhere classified
Information Systems not elsewhere classified
parallel attention mechanism
novel convolutional block
new convolutional block
iou loss function
improved algorithm based
model &# 8217
improved model achieved
varying aspect ratios
experimental results show
depthwise separable convolution
yolov8 </ p
improving feature extraction
blasting sites based
integrate wavelet convolution
average precision increase
separable convolution
blasting sites
model uses
varying scales
results indicate
recall increase
blasting safety
voc2012 dataset
scale convolutions
paper proposes
overlap calculations
object scales
neck network
irregular shapes
effective solution
detonator wires
detect objects
critical issues
complex environment
complex backgrounds
c2f module
built dataset
bounding boxes