BNNDC: Branched neural network for plant disease identification
Deep learning (DL) advancements have contributed to the success of vision-based tasks for solving real-world problems. DL applications in agriculture are increasing as researchers find it valuable for developing solutions to ensure global food security. However, commonly used DL architectures were d...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2023-10-01
|
| Series: | Smart Agricultural Technology |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2772375523001442 |
| _version_ | 1797685253502926848 |
|---|---|
| author | Aanis Ahmad Varun Aggarwal Dharmendra Saraswat |
| author_facet | Aanis Ahmad Varun Aggarwal Dharmendra Saraswat |
| author_sort | Aanis Ahmad |
| collection | DOAJ |
| description | Deep learning (DL) advancements have contributed to the success of vision-based tasks for solving real-world problems. DL applications in agriculture are increasing as researchers find it valuable for developing solutions to ensure global food security. However, commonly used DL architectures were designed for large datasets comprising millions of hand-annotated images. Such large, annotated datasets are limited for agricultural applications. Further, specific agricultural applications, such as disease identification, present a non-trivial problem due to similar visual features of different diseases and background noise. Hence, developing lightweight and task-specific DL architectures may help improve plant disease identification performance. This study proposes a novel neural network architecture named Branched Neural Network for Disease Classification (BNNDC)-Net. The performance of the proposed BNNDC-Net was compared with large (DenseNet169 and ResNet50) and small (ResNet18 and MobileNetV2) conventional DL architectures by training and testing on publicly available and field-acquired data. BNNDC helped eliminate false positives compared to conventional DL models. BNNDC-Net and BNNDC-Res with 3.86M and 6.81M parameters, outperformed ResNet50 for disease identification with 85.60% and 73.36% fewer trainable parameters and 13.93% and 14.26% improvement in testing accuracy, respectively. All trained DL architectures were additionally tested for generalization capability on field images. BNNDC-Res outperformed all other architectures by at least 5.17% in generalization accuracy. This study demonstrates the potential of a lightweight but limited data-requiring DL architecture for plant disease identification and generalization to field conditions. Future work will further expand the study to diagnose diseases for other field crops. |
| first_indexed | 2024-03-12T00:42:27Z |
| format | Article |
| id | doaj.art-af740181894044409a2ad9b387de10ae |
| institution | Directory Open Access Journal |
| issn | 2772-3755 |
| language | English |
| last_indexed | 2024-03-12T00:42:27Z |
| publishDate | 2023-10-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Smart Agricultural Technology |
| spelling | doaj.art-af740181894044409a2ad9b387de10ae2023-09-15T04:40:37ZengElsevierSmart Agricultural Technology2772-37552023-10-015100315BNNDC: Branched neural network for plant disease identificationAanis Ahmad0Varun Aggarwal1Dharmendra Saraswat2Elmore Family School of Electrical and Computer Engineering, Purdue University West Lafayette, IN, United StatesElmore Family School of Electrical and Computer Engineering, Purdue University West Lafayette, IN, United StatesAgricultural and Biological Engineering, Purdue University West Lafayette, IN, United States; Corresponding author.Deep learning (DL) advancements have contributed to the success of vision-based tasks for solving real-world problems. DL applications in agriculture are increasing as researchers find it valuable for developing solutions to ensure global food security. However, commonly used DL architectures were designed for large datasets comprising millions of hand-annotated images. Such large, annotated datasets are limited for agricultural applications. Further, specific agricultural applications, such as disease identification, present a non-trivial problem due to similar visual features of different diseases and background noise. Hence, developing lightweight and task-specific DL architectures may help improve plant disease identification performance. This study proposes a novel neural network architecture named Branched Neural Network for Disease Classification (BNNDC)-Net. The performance of the proposed BNNDC-Net was compared with large (DenseNet169 and ResNet50) and small (ResNet18 and MobileNetV2) conventional DL architectures by training and testing on publicly available and field-acquired data. BNNDC helped eliminate false positives compared to conventional DL models. BNNDC-Net and BNNDC-Res with 3.86M and 6.81M parameters, outperformed ResNet50 for disease identification with 85.60% and 73.36% fewer trainable parameters and 13.93% and 14.26% improvement in testing accuracy, respectively. All trained DL architectures were additionally tested for generalization capability on field images. BNNDC-Res outperformed all other architectures by at least 5.17% in generalization accuracy. This study demonstrates the potential of a lightweight but limited data-requiring DL architecture for plant disease identification and generalization to field conditions. Future work will further expand the study to diagnose diseases for other field crops.http://www.sciencedirect.com/science/article/pii/S2772375523001442Deep learningImage classificationDisease identificationHierarchical architectureBranched architectureGeneralization |
| spellingShingle | Aanis Ahmad Varun Aggarwal Dharmendra Saraswat BNNDC: Branched neural network for plant disease identification Smart Agricultural Technology Deep learning Image classification Disease identification Hierarchical architecture Branched architecture Generalization |
| title | BNNDC: Branched neural network for plant disease identification |
| title_full | BNNDC: Branched neural network for plant disease identification |
| title_fullStr | BNNDC: Branched neural network for plant disease identification |
| title_full_unstemmed | BNNDC: Branched neural network for plant disease identification |
| title_short | BNNDC: Branched neural network for plant disease identification |
| title_sort | bnndc branched neural network for plant disease identification |
| topic | Deep learning Image classification Disease identification Hierarchical architecture Branched architecture Generalization |
| url | http://www.sciencedirect.com/science/article/pii/S2772375523001442 |
| work_keys_str_mv | AT aanisahmad bnndcbranchedneuralnetworkforplantdiseaseidentification AT varunaggarwal bnndcbranchedneuralnetworkforplantdiseaseidentification AT dharmendrasaraswat bnndcbranchedneuralnetworkforplantdiseaseidentification |