Confusion matrix of ResNet50.
<div><p>Accurately recognizing rice seed varieties poses significant challenges due to their diverse morphological characteristics and complex classification requirements. Traditional image recognition methods often struggle with both accuracy and efficiency in this context. To address t...
محفوظ في:
| المؤلف الرئيسي: | |
|---|---|
| مؤلفون آخرون: | , , , , , |
| منشور في: |
2025
|
| الموضوعات: | |
| الوسوم: |
إضافة وسم
لا توجد وسوم, كن أول من يضع وسما على هذه التسجيلة!
|
| _version_ | 1852020310082060288 |
|---|---|
| author | Tingyuan Zhang (21376939) |
| author2 | Changsheng Zhang (1450165) Zhongyi Yang (206009) Meng Wang (124646) Fujie Zhang (155391) Dekai Li (21376942) Sen Yang (88300) |
| author2_role | author author author author author author |
| author_facet | Tingyuan Zhang (21376939) Changsheng Zhang (1450165) Zhongyi Yang (206009) Meng Wang (124646) Fujie Zhang (155391) Dekai Li (21376942) Sen Yang (88300) |
| author_role | author |
| dc.creator.none.fl_str_mv | Tingyuan Zhang (21376939) Changsheng Zhang (1450165) Zhongyi Yang (206009) Meng Wang (124646) Fujie Zhang (155391) Dekai Li (21376942) Sen Yang (88300) |
| dc.date.none.fl_str_mv | 2025-05-16T17:34:48Z |
| dc.identifier.none.fl_str_mv | 10.1371/journal.pone.0322699.g013 |
| dc.relation.none.fl_str_mv | https://figshare.com/articles/figure/Confusion_matrix_of_ResNet50_/29089351 |
| dc.rights.none.fl_str_mv | CC BY 4.0 info:eu-repo/semantics/openAccess |
| dc.subject.none.fl_str_mv | Biotechnology Sociology Space Science Environmental Sciences not elsewhere classified Biological Sciences not elsewhere classified Chemical Sciences not elsewhere classified Information Systems not elsewhere classified rice seed classification minimizing redundant features global receptive field four convolutional stages double attention mechanism diverse morphological characteristics complex classification requirements collected dataset demonstrate addressing similar fine 94 %, surpassing grained recognition challenges channel residual network net architecture consists div >< p 2 </ sup 22 %, respectively 72 %, network ’ recognition accuracy swin transformer superior performance study proposes promising reference net provides net achieves improves inter findings confirm experimental results efficient solution core innovation class differentiation block ), baseline model art models 16 layers 16 %. |
| dc.title.none.fl_str_mv | Confusion matrix of ResNet50. |
| dc.type.none.fl_str_mv | Image Figure info:eu-repo/semantics/publishedVersion image |
| description | <div><p>Accurately recognizing rice seed varieties poses significant challenges due to their diverse morphological characteristics and complex classification requirements. Traditional image recognition methods often struggle with both accuracy and efficiency in this context. To address these limitations, this study proposes the Deep Space and Channel Residual Network with Double Attention Mechanism (RSCD-Net) to enhance the recognition accuracy of 36 rice seed varieties. The core innovation of RSCD-Net is the introduction of the Space and Channel Feature Extraction Residual Block (SCR-Block), which improves inter-class differentiation while minimizing redundant features, thereby optimizing computational efficiency. The RSCD-Net architecture consists of 16 layers of SCR-Blocks, structured into four convolutional stages with 3, 4, 6, and 3 units, respectively. Additionally, a Double Attention Mechanism (A<sup>2</sup>Net) is incorporated to enhance the network’s global receptive field, improving its capacity to distinguish subtle variations among seed types. Experimental results on a self-collected dataset demonstrate that RSCD-Net achieves an average accuracy of 81.94%, surpassing the baseline model by 4.16%. Compared with state-of-the-art models such as InceptionResNetV2, ConvNeXt, MobileNetV3, and Swin Transformer, RSCD Net has improved by 1.17%, 3%, 24.72%, and 13.22%, respectively, showcasing its superior performance. These findings confirm that RSCD-Net provides an effective and efficient solution for rice seed classification, offering a promising reference for addressing similar fine-grained recognition challenges in agricultural applications.</p></div> |
| eu_rights_str_mv | openAccess |
| id | Manara_7c2fa2bfe95208c8a1bcce347ee9e425 |
| identifier_str_mv | 10.1371/journal.pone.0322699.g013 |
| network_acronym_str | Manara |
| network_name_str | ManaraRepo |
| oai_identifier_str | oai:figshare.com:article/29089351 |
| publishDate | 2025 |
| repository.mail.fl_str_mv | |
| repository.name.fl_str_mv | |
| repository_id_str | |
| rights_invalid_str_mv | CC BY 4.0 |
| spelling | Confusion matrix of ResNet50.Tingyuan Zhang (21376939)Changsheng Zhang (1450165)Zhongyi Yang (206009)Meng Wang (124646)Fujie Zhang (155391)Dekai Li (21376942)Sen Yang (88300)BiotechnologySociologySpace ScienceEnvironmental Sciences not elsewhere classifiedBiological Sciences not elsewhere classifiedChemical Sciences not elsewhere classifiedInformation Systems not elsewhere classifiedrice seed classificationminimizing redundant featuresglobal receptive fieldfour convolutional stagesdouble attention mechanismdiverse morphological characteristicscomplex classification requirementscollected dataset demonstrateaddressing similar fine94 %, surpassinggrained recognition challengeschannel residual networknet architecture consistsdiv >< p2 </ sup22 %, respectively72 %,network ’recognition accuracyswin transformersuperior performancestudy proposespromising referencenet providesnet achievesimproves interfindings confirmexperimental resultsefficient solutioncore innovationclass differentiationblock ),baseline modelart models16 layers16 %.<div><p>Accurately recognizing rice seed varieties poses significant challenges due to their diverse morphological characteristics and complex classification requirements. Traditional image recognition methods often struggle with both accuracy and efficiency in this context. To address these limitations, this study proposes the Deep Space and Channel Residual Network with Double Attention Mechanism (RSCD-Net) to enhance the recognition accuracy of 36 rice seed varieties. The core innovation of RSCD-Net is the introduction of the Space and Channel Feature Extraction Residual Block (SCR-Block), which improves inter-class differentiation while minimizing redundant features, thereby optimizing computational efficiency. The RSCD-Net architecture consists of 16 layers of SCR-Blocks, structured into four convolutional stages with 3, 4, 6, and 3 units, respectively. Additionally, a Double Attention Mechanism (A<sup>2</sup>Net) is incorporated to enhance the network’s global receptive field, improving its capacity to distinguish subtle variations among seed types. Experimental results on a self-collected dataset demonstrate that RSCD-Net achieves an average accuracy of 81.94%, surpassing the baseline model by 4.16%. Compared with state-of-the-art models such as InceptionResNetV2, ConvNeXt, MobileNetV3, and Swin Transformer, RSCD Net has improved by 1.17%, 3%, 24.72%, and 13.22%, respectively, showcasing its superior performance. These findings confirm that RSCD-Net provides an effective and efficient solution for rice seed classification, offering a promising reference for addressing similar fine-grained recognition challenges in agricultural applications.</p></div>2025-05-16T17:34:48ZImageFigureinfo:eu-repo/semantics/publishedVersionimage10.1371/journal.pone.0322699.g013https://figshare.com/articles/figure/Confusion_matrix_of_ResNet50_/29089351CC BY 4.0info:eu-repo/semantics/openAccessoai:figshare.com:article/290893512025-05-16T17:34:48Z |
| spellingShingle | Confusion matrix of ResNet50. Tingyuan Zhang (21376939) Biotechnology Sociology Space Science Environmental Sciences not elsewhere classified Biological Sciences not elsewhere classified Chemical Sciences not elsewhere classified Information Systems not elsewhere classified rice seed classification minimizing redundant features global receptive field four convolutional stages double attention mechanism diverse morphological characteristics complex classification requirements collected dataset demonstrate addressing similar fine 94 %, surpassing grained recognition challenges channel residual network net architecture consists div >< p 2 </ sup 22 %, respectively 72 %, network ’ recognition accuracy swin transformer superior performance study proposes promising reference net provides net achieves improves inter findings confirm experimental results efficient solution core innovation class differentiation block ), baseline model art models 16 layers 16 %. |
| status_str | publishedVersion |
| title | Confusion matrix of ResNet50. |
| title_full | Confusion matrix of ResNet50. |
| title_fullStr | Confusion matrix of ResNet50. |
| title_full_unstemmed | Confusion matrix of ResNet50. |
| title_short | Confusion matrix of ResNet50. |
| title_sort | Confusion matrix of ResNet50. |
| topic | Biotechnology Sociology Space Science Environmental Sciences not elsewhere classified Biological Sciences not elsewhere classified Chemical Sciences not elsewhere classified Information Systems not elsewhere classified rice seed classification minimizing redundant features global receptive field four convolutional stages double attention mechanism diverse morphological characteristics complex classification requirements collected dataset demonstrate addressing similar fine 94 %, surpassing grained recognition challenges channel residual network net architecture consists div >< p 2 </ sup 22 %, respectively 72 %, network ’ recognition accuracy swin transformer superior performance study proposes promising reference net provides net achieves improves inter findings confirm experimental results efficient solution core innovation class differentiation block ), baseline model art models 16 layers 16 %. |