Multifunctional cells based neural architecture search for plant images classification
Abstract To develop a high-performance convolutional neural network (CNN) model for plant image classification automatically, we propose a neural architecture search (NAS) method tailored to multifunctional cells (MFC), termed MFC-NAS. Initially, a search space based on MFC is designed, encompassing...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-07-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-11829-7 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849766707693879296 |
|---|---|
| author | Lin Huang Xi Qin Tiejun Yang |
| author_facet | Lin Huang Xi Qin Tiejun Yang |
| author_sort | Lin Huang |
| collection | DOAJ |
| description | Abstract To develop a high-performance convolutional neural network (CNN) model for plant image classification automatically, we propose a neural architecture search (NAS) method tailored to multifunctional cells (MFC), termed MFC-NAS. Initially, a search space based on MFC is designed, encompassing transfer cell, normal cell, pooling cell, and dropout cell, with transfer cell dedicated to exploring weight-sharing layers. Subsequently, an MFC-oriented search strategy is adopted: different shallow blocks from pre-trained models such as MobileNet V3 are searched to construct transfer cell. Similar strategies are applied to pooling cell, dropout cell, and normal cell, exploring diverse pooling types and sizes for pooling cell and various dropout rates for dropout cell. Finally, the best-found cells are stacked to form a plant image classification CNN based on MFC. Experiments conducted on two publicly available plant image datasets demonstrate that MFC-NAS achieves the optimal cells after approximately 69 GPU-hours of search. Compared to state-of-the-art (SOTA) methods like ResNet-50 and EfficientNet, this approach attains higher accuracy (~ 99.10%) with an average single-sample inference time of around 12.6 ms. Moreover, the number of network parameters used in the proposed method is only 6.9% of ResNet-50’s (approximately 1.58 M). |
| format | Article |
| id | doaj-art-c8b08f2dd10a480eab5081a93c4e9be9 |
| institution | DOAJ |
| issn | 2045-2322 |
| language | English |
| publishDate | 2025-07-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj-art-c8b08f2dd10a480eab5081a93c4e9be92025-08-20T03:04:29ZengNature PortfolioScientific Reports2045-23222025-07-0115111410.1038/s41598-025-11829-7Multifunctional cells based neural architecture search for plant images classificationLin Huang0Xi Qin1Tiejun Yang2College of Physics and Electronic Information Engineering, Guilin University of TechnologyCollege of Physics and Electronic Information Engineering, Guilin University of TechnologyCollege of Intelligent Medicine and Biotechnology, Guilin Medical UniversityAbstract To develop a high-performance convolutional neural network (CNN) model for plant image classification automatically, we propose a neural architecture search (NAS) method tailored to multifunctional cells (MFC), termed MFC-NAS. Initially, a search space based on MFC is designed, encompassing transfer cell, normal cell, pooling cell, and dropout cell, with transfer cell dedicated to exploring weight-sharing layers. Subsequently, an MFC-oriented search strategy is adopted: different shallow blocks from pre-trained models such as MobileNet V3 are searched to construct transfer cell. Similar strategies are applied to pooling cell, dropout cell, and normal cell, exploring diverse pooling types and sizes for pooling cell and various dropout rates for dropout cell. Finally, the best-found cells are stacked to form a plant image classification CNN based on MFC. Experiments conducted on two publicly available plant image datasets demonstrate that MFC-NAS achieves the optimal cells after approximately 69 GPU-hours of search. Compared to state-of-the-art (SOTA) methods like ResNet-50 and EfficientNet, this approach attains higher accuracy (~ 99.10%) with an average single-sample inference time of around 12.6 ms. Moreover, the number of network parameters used in the proposed method is only 6.9% of ResNet-50’s (approximately 1.58 M).https://doi.org/10.1038/s41598-025-11829-7Neural architecture searchMultifunctional cellsTransfer cellDeep learningPlant images classification |
| spellingShingle | Lin Huang Xi Qin Tiejun Yang Multifunctional cells based neural architecture search for plant images classification Scientific Reports Neural architecture search Multifunctional cells Transfer cell Deep learning Plant images classification |
| title | Multifunctional cells based neural architecture search for plant images classification |
| title_full | Multifunctional cells based neural architecture search for plant images classification |
| title_fullStr | Multifunctional cells based neural architecture search for plant images classification |
| title_full_unstemmed | Multifunctional cells based neural architecture search for plant images classification |
| title_short | Multifunctional cells based neural architecture search for plant images classification |
| title_sort | multifunctional cells based neural architecture search for plant images classification |
| topic | Neural architecture search Multifunctional cells Transfer cell Deep learning Plant images classification |
| url | https://doi.org/10.1038/s41598-025-11829-7 |
| work_keys_str_mv | AT linhuang multifunctionalcellsbasedneuralarchitecturesearchforplantimagesclassification AT xiqin multifunctionalcellsbasedneuralarchitecturesearchforplantimagesclassification AT tiejunyang multifunctionalcellsbasedneuralarchitecturesearchforplantimagesclassification |