MFDPN module.
<div><p>Personal protective equipment (PPE) is critical for ensuring the safety of construction workers. However, site surveillance images from construction sites often feature multi-size and multi-scale targets, leading to low detection accuracy for PPE in existing models. To address th...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | , , , |
| Published: |
2025
|
| Subjects: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1852019903368790016 |
|---|---|
| author | Bo Tong (2138632) |
| author2 | Guan Li (3820189) Xiangli Bu (771270) Yang Wang (5921) Xingchen Yu (8933981) |
| author2_role | author author author author |
| author_facet | Bo Tong (2138632) Guan Li (3820189) Xiangli Bu (771270) Yang Wang (5921) Xingchen Yu (8933981) |
| author_role | author |
| dc.creator.none.fl_str_mv | Bo Tong (2138632) Guan Li (3820189) Xiangli Bu (771270) Yang Wang (5921) Xingchen Yu (8933981) |
| dc.date.none.fl_str_mv | 2025-05-29T17:27:38Z |
| dc.identifier.none.fl_str_mv | 10.1371/journal.pone.0322115.g003 |
| dc.relation.none.fl_str_mv | https://figshare.com/articles/figure/MFDPN_module_/29186678 |
| dc.rights.none.fl_str_mv | CC BY 4.0 info:eu-repo/semantics/openAccess |
| dc.subject.none.fl_str_mv | Space Science Biological Sciences not elsewhere classified Information Systems not elsewhere classified site surveillance images image data distribution generating spatial offsets experimental results indicate dynamically selecting weights detailed contextual information 6 %, respectively multiscale feature focus enriching feature diversity structured pruning techniques computational load decreased model &# 8217 enhance classification precision parameter count dropped integrate interactive features low detection accuracy improved model based parameter count interactive features feature masks reducing computational detection accuracy based algorithm parameter loads varying levels substantial reduction scale features pruning level proposed mfd paper proposes localization tasks geometric transformations existing models deep learning construction workers construction sites construction scenarios better adapt across scales 95 improved 53 %. |
| dc.title.none.fl_str_mv | MFDPN module. |
| dc.type.none.fl_str_mv | Image Figure info:eu-repo/semantics/publishedVersion image |
| description | <div><p>Personal protective equipment (PPE) is critical for ensuring the safety of construction workers. However, site surveillance images from construction sites often feature multi-size and multi-scale targets, leading to low detection accuracy for PPE in existing models. To address this issue, this paper proposes an improved model based on YOLOv8n.By enriching feature diversity and enhancing the model’s adaptability to geometric transformations, the detection accuracy is improved.A Multi-Scale Group Convolution Module (MSGP) was designed to extract multi-level features using different convolution kernels. A Multi-Scale Feature Diffusion Pyramid Network (MFDPN) was developed, which aggregates multi-scale features through the Multiscale Feature Focus (MFF) module and diffuses them across scales, providing each scale with detailed contextual information. A customized Task Alignment Module was introduced to integrate interactive features, optimizing both classification and localization tasks. The DCNV2(Deformable Convolutional Networks v2) module was incorporated to handle geometric scale transformations by generating spatial offsets and feature masks from interactive features, thereby improving localization accuracy and dynamically selecting weights to enhance classification precision.The improved model incorporates rich multi-level and multi-scale features, allowing it to better adapt to tasks involving geometric transformations and aligning with the image data distribution in construction scenarios. Additionally, structured pruning techniques were applied to the model at varying levels, further reducing computational and parameter loads. Experimental results indicate that at a pruning level of 1.5, mAP@0.5 and mAP@0.5:0.95 improved by 3.9% and 4.6%, respectively, while computational load decreased by 21% and parameter count dropped by 53%. The proposed MFD-YOLO(1.5) model achieves significant progress in detecting personal protective equipment on construction sites, with a substantial reduction in parameter count, making it suitable for deployment on resource-constrained edge devices.</p></div> |
| eu_rights_str_mv | openAccess |
| id | Manara_fa2a42b375ea2ab1046bf90ff3f24f76 |
| identifier_str_mv | 10.1371/journal.pone.0322115.g003 |
| network_acronym_str | Manara |
| network_name_str | ManaraRepo |
| oai_identifier_str | oai:figshare.com:article/29186678 |
| publishDate | 2025 |
| repository.mail.fl_str_mv | |
| repository.name.fl_str_mv | |
| repository_id_str | |
| rights_invalid_str_mv | CC BY 4.0 |
| spelling | MFDPN module.Bo Tong (2138632)Guan Li (3820189)Xiangli Bu (771270)Yang Wang (5921)Xingchen Yu (8933981)Space ScienceBiological Sciences not elsewhere classifiedInformation Systems not elsewhere classifiedsite surveillance imagesimage data distributiongenerating spatial offsetsexperimental results indicatedynamically selecting weightsdetailed contextual information6 %, respectivelymultiscale feature focusenriching feature diversitystructured pruning techniquescomputational load decreasedmodel &# 8217enhance classification precisionparameter count droppedintegrate interactive featureslow detection accuracyimproved model basedparameter countinteractive featuresfeature masksreducing computationaldetection accuracybased algorithmparameter loadsvarying levelssubstantial reductionscale featurespruning levelproposed mfdpaper proposeslocalization tasksgeometric transformationsexisting modelsdeep learningconstruction workersconstruction sitesconstruction scenariosbetter adaptacross scales95 improved53 %.<div><p>Personal protective equipment (PPE) is critical for ensuring the safety of construction workers. However, site surveillance images from construction sites often feature multi-size and multi-scale targets, leading to low detection accuracy for PPE in existing models. To address this issue, this paper proposes an improved model based on YOLOv8n.By enriching feature diversity and enhancing the model’s adaptability to geometric transformations, the detection accuracy is improved.A Multi-Scale Group Convolution Module (MSGP) was designed to extract multi-level features using different convolution kernels. A Multi-Scale Feature Diffusion Pyramid Network (MFDPN) was developed, which aggregates multi-scale features through the Multiscale Feature Focus (MFF) module and diffuses them across scales, providing each scale with detailed contextual information. A customized Task Alignment Module was introduced to integrate interactive features, optimizing both classification and localization tasks. The DCNV2(Deformable Convolutional Networks v2) module was incorporated to handle geometric scale transformations by generating spatial offsets and feature masks from interactive features, thereby improving localization accuracy and dynamically selecting weights to enhance classification precision.The improved model incorporates rich multi-level and multi-scale features, allowing it to better adapt to tasks involving geometric transformations and aligning with the image data distribution in construction scenarios. Additionally, structured pruning techniques were applied to the model at varying levels, further reducing computational and parameter loads. Experimental results indicate that at a pruning level of 1.5, mAP@0.5 and mAP@0.5:0.95 improved by 3.9% and 4.6%, respectively, while computational load decreased by 21% and parameter count dropped by 53%. The proposed MFD-YOLO(1.5) model achieves significant progress in detecting personal protective equipment on construction sites, with a substantial reduction in parameter count, making it suitable for deployment on resource-constrained edge devices.</p></div>2025-05-29T17:27:38ZImageFigureinfo:eu-repo/semantics/publishedVersionimage10.1371/journal.pone.0322115.g003https://figshare.com/articles/figure/MFDPN_module_/29186678CC BY 4.0info:eu-repo/semantics/openAccessoai:figshare.com:article/291866782025-05-29T17:27:38Z |
| spellingShingle | MFDPN module. Bo Tong (2138632) Space Science Biological Sciences not elsewhere classified Information Systems not elsewhere classified site surveillance images image data distribution generating spatial offsets experimental results indicate dynamically selecting weights detailed contextual information 6 %, respectively multiscale feature focus enriching feature diversity structured pruning techniques computational load decreased model &# 8217 enhance classification precision parameter count dropped integrate interactive features low detection accuracy improved model based parameter count interactive features feature masks reducing computational detection accuracy based algorithm parameter loads varying levels substantial reduction scale features pruning level proposed mfd paper proposes localization tasks geometric transformations existing models deep learning construction workers construction sites construction scenarios better adapt across scales 95 improved 53 %. |
| status_str | publishedVersion |
| title | MFDPN module. |
| title_full | MFDPN module. |
| title_fullStr | MFDPN module. |
| title_full_unstemmed | MFDPN module. |
| title_short | MFDPN module. |
| title_sort | MFDPN module. |
| topic | Space Science Biological Sciences not elsewhere classified Information Systems not elsewhere classified site surveillance images image data distribution generating spatial offsets experimental results indicate dynamically selecting weights detailed contextual information 6 %, respectively multiscale feature focus enriching feature diversity structured pruning techniques computational load decreased model &# 8217 enhance classification precision parameter count dropped integrate interactive features low detection accuracy improved model based parameter count interactive features feature masks reducing computational detection accuracy based algorithm parameter loads varying levels substantial reduction scale features pruning level proposed mfd paper proposes localization tasks geometric transformations existing models deep learning construction workers construction sites construction scenarios better adapt across scales 95 improved 53 %. |