CNDD-Net: A Lightweight Attention-Based Convolutional Neural Network for Classifying Corn Nutritional Deficiencies and Leaf Diseases
Abstract
:1. Introduction
2. Materials and Methods
2.1. Data Collection and Preprocessing
2.2. Lightweight Attention-Based Architecture Design
2.3. Depth-Wise Separable Convolution (DSC)
2.4. Convolution Block Attention Mechanism (CBAM)
3. Experiments and Discussion
3.1. Experimental Conditions
3.2. Experiment 1: With Variation in Channel Numbers (Feature Maps)
3.3. Experiment 2: With Variation in Layer Number (Depth of Network)
3.4. Experiment 3: With Variation in Attention Block Position
3.5. Evaluation Metrics of Lightweight CNDD-Net
3.6. Performance Evaluation of This Research with Conventional Model and Previous Studies
4. Conclusions
Author Contributions
Funding
Data Availability Statement
Acknowledgments
Conflicts of Interest
Abbreviations
CBAM | Convolution Block Attention Module |
CI | Confidence Interval |
CNDD-Net | Corn Leaf Nutrition Deficiency and Disease Network |
CNN | Convolutional Neural Network |
DSC | Depth-wise Separable Convolution |
FC | Fully Connected |
FLOPs | Floating Point Operations Per Second |
FN | False Negative |
FNR | False Negative Rate |
FP | False Positive |
FPR | False Positive Rate |
GFLOPs | Giga Floating Point Operations Per Second |
MLN | Maize Lethal Necrosis |
MSV | Maize Streak Virus |
ReLU | Rectified Linear Unit |
SVM | Support Vector Machine |
TN | True Negative |
TP | True Positive |
References
- Rouf, S.T.; Prasad, K.; Kumar, P.; Yildiz, F. Maize-A potential source of human nutrition and health: A review. In Contaminants in Agriculture; Springer: Berlin/Heidelberg, Germany, 2016; Volume 2. [Google Scholar] [CrossRef]
- Sapna; Archna, J.; Dar, Z.A.; Qureshi, A.M.I. Maize Fabric: Leading to a New Paradigm Shift. IJCMAS 2020, 9, 1337–1343. [Google Scholar] [CrossRef]
- FAO. World Food and Agriculture—Statistical Yearbook 2023; FAO: Rome, Italy, 2023. [Google Scholar] [CrossRef]
- OECD. OECD-FAO Agricultural Outlook 2023–2032; OECD: Paris, France, 2023. [Google Scholar] [CrossRef]
- Boddupalli, P.; Suresh, L.M.; Mwatuni, F.; Beyene, Y.; Makumbi, D.; Gowda, M.; Olsen, M.; Hodson, D.; Worku, M.; Mezzalama, M.; et al. Maize lethal necrosis (MLN): Efforts toward containing the spread and impact of a devastating transboundary disease in sub-Saharan Africa. Virus Res. 2020, 282, 197943. [Google Scholar] [CrossRef]
- Haque, M.A.; Marwaha, S.; Deb, C.K.; Nigam, S.; Arora, A.; Hooda, K.S.; Soujanya, L.; Aggarwal, S.K.; Lall, B.; Kumar, M.; et al. Deep learning-based approach for identification of diseases of maize crop. Sci. Rep. 2022, 12, 6334. [Google Scholar] [CrossRef]
- Lucas, R.E.; Holtman, J.B.; Connor, L.J. Soil carbon dynamics and cropping practices. In Agriculture and Energy; Lockeretz, W., Ed.; Academic Press: New York, NY, USA, 1977; pp. 333–351. [Google Scholar] [CrossRef]
- Leena, N.; Saju, K.K. Classification of macronutrient deficiencies in maize plants using optimized multi class support vector machines. Eng. Agric. Environ. Food 2019, 12, 126–139. [Google Scholar] [CrossRef]
- Simhadri, C.G.; Kondaveeti, H.K.; Vatsavayi, V.K.; Mitra, A.; Ananthachari, P. Deep learning for rice leaf disease detection: A systematic literature review on emerging trends, methodologies and techniques. Inf. Process Agric. 2024, in press. [Google Scholar] [CrossRef]
- CIMMYT Maize Viral Diseases: MLN and MSV. CIMMYT. Available online: https://repository.cimmyt.org/server/api/core/bitstreams/7a4fdc33-0344-4055-99e1-37e226360605/content (accessed on 5 February 2025).
- Kaya, Y.; Gursoy, E. A novel multi-head CNN design to identify plant diseases using the fusion of RGB images. Ecol. Inform. 2023, 75, 101998. [Google Scholar] [CrossRef]
- Patil, M.A.; Manur, M. Sensitive crop leaf disease prediction based on computer vision techniques with handcrafted features. Int. J. Syst. Assur. Eng. Manag. 2023, 14, 2235–2266. [Google Scholar] [CrossRef]
- Khan, F.; Zafar, N.; Tahir, M.N.; Aqib, M.; Waheed, H.; Haroon, Z. A mobile-based system for maize plant leaf disease detection and classification using deep learning. Front. Plant Sci. 2023, 14, 1079366. [Google Scholar] [CrossRef]
- Nan, F.; Song, Y.; Yu, X.; Nie, C.; Liu, Y.; Bai, Y.; Zou, D.; Wang, C.; Yin, D.; Yang, W.; et al. A novel method for maize leaf disease classification using the RGB-D post-segmentation image data. Front. Plant Sci. 2023, 14, 1268015. [Google Scholar] [CrossRef]
- Pratama, A.; Pristyanto, Y. Classification of corn plant diseases using various convolutional neural network. JITK 2023, 9, 49–56. [Google Scholar] [CrossRef]
- Zeng, W.; Li, H.; Hu, G.; Liang, D. Lightweight dense-scale network (LDSNet) for corn leaf disease identification. Comput. Electron. Agric. 2022, 197, 106943. [Google Scholar] [CrossRef]
- Fan, X.; Guan, Z. VGNET: A Lightweight Intelligent Learning Method for corn Diseases Recognition. Agriculture 2023, 13, 1606. [Google Scholar] [CrossRef]
- Simonyan, K.; Zisserman, A. Very deep convolutional networks for large-scale image recognition. arXiv 2015, arXiv:1409.1556. [Google Scholar]
- Thakur, P.S.; Sheorey, T.; Ojha, A. VGG-ICNN: A Lightweight CNN model for crop disease identification. Multimed. Tools Appl. 2023, 82, 497–520. [Google Scholar] [CrossRef]
- Szegedy, C.; Liu, W.; Jia, Y.; Sermanet, P.; Reed, S.; Anguelov, D.; Erhan, D.; Vanhoucke, V.; Rabinovich, A. Going deeper with convolutions. In Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Boston, MA, USA, 7–12 June 2015; pp. 1–9. [Google Scholar] [CrossRef]
- Bera, A.; Bhattacharjee, D.; Krejcar, O. PND-net: Plant Nutrition deficiency and disease classification using graph Convolutional Network. Sci. Rep. 2024, 14, 1079366. [Google Scholar] [CrossRef]
- Ibrahim, S.; Hasan, N.; Sabri, N.; Abu Samah, K.; Rahimi Rusland, M. Palm leaf nutrient deficiency detection using convolutional neural network (CNN). IJNAA 2022, 13, 1949–1956. [Google Scholar] [CrossRef]
- Parnal, P. Pawade. Harnessing Deep Learning for Plant Nutrition: VGG Architecture for Precision Detection of Nutrient Deficiency. IJISAE 2024, 12, 2757–2766. [Google Scholar]
- Wang, C.; Ye, Y.; Tian, Y.; Yu, Z. Classification of nutrient deficiency in rice based on CNN model with Reinforcement Learning Augmentation. In Proceedings of the 2021 International Symposium on Artificial Intelligence and Its Application on Media (ISAIAM), Xi’an, China, 21–23 May 2021. [Google Scholar] [CrossRef]
- Ramos-Ospina, M.; Gomez, L.; Trujillo, C.; Marulanda-Tob’on, A. Deep Transfer Learning for Image Classification of Phosphorus Nutrition States in Individual Maize Leaves. Electronics 2024, 13, 16. [Google Scholar] [CrossRef]
- Bhujel, A.; Kim, N.-E.; Arulmozhi, E.; Basak, J.K.; Kim, H.-T. A Lightweight Attention Based Convolutional Neural Networks for Tomato Leaf Disease Classification. Agriculture 2022, 12, 228. [Google Scholar] [CrossRef]
- Mnih, V.; Heess, N.; Graves, A.; Kavukcuoglu, K. Recurrent models of visual attention. In Proceedings of the Advances in Neural Information Processing Systems: Annual Conference on Neural Information Processing Systems 2014, Montreal, QC, Canada, 8–13 December 2014; Volume 3, pp. 2204–2212. [Google Scholar]
- Woo, S.; Park, J.; Lee, J.Y.; Kweon, I.S. CBAM: Convolutional block attention module. In Proceedings of the Computer Vision—ECCV 2018, Munich, Germany, 8–14 September 2018; Volume 11211, pp. 3–19. [Google Scholar] [CrossRef]
- Zeng, W.; Li, M. Crop leaf disease recognition based on Self-Attention convolutional neural network. Comput. Electron. Agric. 2020, 172, 105341. [Google Scholar]
- Lee, S.H.; Goëau, H.; Bonnet, P.; Joly, A. Attention-Based Recurrent Neural Network for Plant Disease Classification. Front. Plant Sci. 2020, 11, 601250. [Google Scholar] [CrossRef] [PubMed]
- He, K.; Zhang, X.; Ren, S.; Sun, J. Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, 10–17 June 2016; pp. 27–30. [Google Scholar] [CrossRef]
- Shrivastav, P.; Prasad, M.; Singh, T.B.; Yadav, A.; Goyal, D.; Ali, A.; Dantu, P.K. Role of Nutrients in Plant Growth and Development. In Contaminants in Agriculture; Naeem, M., Ansari, A., Gill, S., Eds.; Springer: Cham, Switzerland, 2020. [Google Scholar] [CrossRef]
- Nikitha, S.; Prabhanjan, S.; Sathyanarayan, A. Plant nutritional deficiency detection: A survey of predictive analytics approaches. Iran J. Comput. Sci. 2024, 8, 83–101. [Google Scholar] [CrossRef]
- Missouri Integrated Pest Management. Diagnosing Nutrient Deficiencies. University of Missouri. Division of Plant Sciences. 2011. Available online: https://ipm.missouri.edu/MEG/2011/6/Diagnosing-Nutrient-Deficiencies/ (accessed on 7 January 2025).
- Jung, M.; Song, J.S.; Shin, A.; Choi, B.; Go, S.; Kwon, S.; Park, J.; Park, S.G.; Kim, Y. Construction of deep learning-based disease detection model in plants. Sci. Rep. 2023, 13, 7331. [Google Scholar] [CrossRef]
- Sudhakar, M.; Priya, R.M.S. Computer vision-based machine learning and deep learning approaches for identification of nutrient deficiency in crops: A survey. NEPT 2023, 22, 1387–1399. [Google Scholar] [CrossRef]
- Mduma, N.; Laizer, H. Machine learning imagery dataset for maize crop: A case of Tanzania. Data Brief. 2023, 48, 109108. [Google Scholar] [CrossRef]
- Climate-Data. Climate data for cities worldwide—Climate-Data. 2019. Available online: https://en.climate-data.org (accessed on 27 March 2025).
- Supani, A.; Andriani, Y.; Indarto; Saputra, H.; Joni, A.B.; Alfian, D.; Taqwa, A.; Silvia, A.H. Enhancing deeper layers with residual network on CNN Architecture: A Review. AHE 2023, 14, 449–457. [Google Scholar] [CrossRef]
- Zhou, X.; Liu, H.; Shi, C.; Liu, J. Deep Learning on Edge Computing Devices: Design Challenges of Algorithm and Architecture; Elsevier: Amsterdam, The Netherlands, 2022. [Google Scholar] [CrossRef]
- Howard, A.G.; Zhu, M.; Chen, B.; Kalenichenko, D.; Wang, W.; Weyand, T.; Andreetto, M.; Adam, H. MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications. arXiv 2017, arXiv:1704.04861. [Google Scholar]
- Tan, M.; Le, Q.V. EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks. In Proceedings of the 36th International Conference on Machine Learning, ICML 2019, Long Beach, CA, USA, 9–15 June 2019; pp. 6105–6114. Available online: http://proceedings.mlr.press/v97/tan19a.html (accessed on 15 February 2025).
- Hu, J.; Shen, L.; Albanie, S.; Sun, G.; Wu, E. Squeeze-and-Excitation networks. IEEE Trans. Pattern Anal. Mach. Intell. 2019, 42, 2011–2023. [Google Scholar] [CrossRef]
- Ippolito, P.P. Hyperparameter Tuning. In Applied Data Science in Tourism.TV; Egger, R., Ed.; Springer: Cham, Switzerland, 2022. [Google Scholar] [CrossRef]
- Kingma, D.P.; Ba, J.L. Adam: A method for stochastic optimization. arXiv 2014, arXiv:1412.6980. [Google Scholar]
- Mao, A.; Mohri, M.; Zhong, Y. Cross-Entropy Loss Functions: Theoretical analysis and applications. arXiv 2023, arXiv:2304.07288. [Google Scholar]
- Grandini, M.; Bagli, E.; Visani, G. Metrics for Multi-Class Classification: An Overview. arXiv 2020, arXiv:2008.05756. [Google Scholar]
- Molchanov, P.; Tyree, S.; Karras, T.; Aila, T.; Kautz, J. Pruning Convolutional Neural Networks for Resource Efficient Inference. International Conference on Learning Representations (ICLR). 2016. Available online: https://openreview.net/pdf?id=SJGCiw5gl (accessed on 19 February 2025).
- Programmer Sought. CNN Model Complexity (FLOPS, MAC), Parameter Amount and Running Speed. Available online: https://www.programmersought.com/article/50435353828/ (accessed on 25 February 2025).
- Zhu, H.; An, Z.; Yang, C.; Hu, X.; Xu, K.; Xu, Y. Rethinking the Number of Channels for the Convolutional Neural Network. arXiv 2019, arXiv:1909.01861. [Google Scholar]
- Zhao, X.; Wang, L.; Zhang, Y.; Han, X.; Deveci, M.; Parmar, M. A review of convolutional neural networks in computer vision. Artif. Intell. Rev. 2024, 57, 99. [Google Scholar] [CrossRef]
- Zou, J.; Rui, T.; Zhou, Y.; Yang, C.; Zhang, S. Convolutional neural network simplification via feature map pruning. Comput. Electron. Agric. 2018, 70, 950–958. [Google Scholar] [CrossRef]
- Alzubaidi, L.; Zhang, J.; Humaidi, A.J.; Al-Dujaili, A.; Duan, Y.; Al-Shamma, O.; Santamaría, J.; Fadhel, M.A.; Al-Amidie, M.; Farhan, L. Review of deep learning: Concepts, CNN architectures, challenges, applications, future directions. J. Big Data. 2021, 8, 53. [Google Scholar] [CrossRef]
- Chen, F.; Tsou, J.Y. Assessing the effects of convolutional neural network architectural factors on model performance for Remote Sensing Image Classification: An in-depth investigation. Int. J. Appl Earth Obs. Geoinf. 2022, 112, 102865. [Google Scholar] [CrossRef]
- Hassanin, M.; Anwar, S.; Radwan, I.; Khan, F.S.; Mian, A. Visual attention methods in deep learning: An in-depth survey. Inf. Fusion. 2024, 108, 102417. [Google Scholar] [CrossRef]
- Abbas, S.; Hamouma, M.; Abbas, F. Efficient Method Using Attention Based Convolutional Neural Networks for Ceramic Tiles Defect Classification. Rev. D’intelligence Artif. 2023, 37, 53–62. [Google Scholar] [CrossRef]
- Iandola, F.; Han, S.; Moskewicz, M.; Ashraf, K.; Dally, W.; Keutzer, K. SqueezeNet: AlexNet level accuracy with 50x fewer parameters and <0.5 MB model size. arXiv 2016, arXiv:1602.07360. [Google Scholar]
- Chen, J.; Wang, W.; Zhang, D.; Zeb, A.; Nanehkaran, Y.A. Attention embedded lightweight network for maize disease recognition. Plant Pathol. J. 2020, 70, 630–642. [Google Scholar] [CrossRef]
- Liu, L.; Qiao, S.; Chang, J.; Ding, W.; Xu, C.; Gu, J.; Sun, T.; Qiao, H. A multi-scale feature fusion neural network for multi-class disease classification on the maize leaf images. Heliyon. 2024, 10, e28264. [Google Scholar] [CrossRef] [PubMed]
- Alloway, B.J. Micronutrient Deficiencies in Global Crop Production; Springer eBooks: Berlin/Heidelberg, Germany, 2008. [Google Scholar] [CrossRef]
Class Label | Class Name | Number of Images | Source |
---|---|---|---|
0 | Normal | 1028 | Field |
1 | Iron deficiency | 589 | Field |
2 | Zinc deficiency | 753 | Field |
3 | Phosphorus deficiency | 479 | Field |
4 | Nitrogen deficiency | 372 | Field |
5 | Maize Streak Virus (MSV) | 590 | Public |
6 | Maize Lethal Necrosis (MLN) | 675 | Public |
Total | 4486 |
Layer Name | Output Size | Configuration | Remarks |
---|---|---|---|
Input Image | 224 × 224 × 224 | ||
conv1 | 112 × 112 × 16 | 7 × 7, 16, Stride 2 3 × 3 max pool, Stride 2 | |
Conv2_x | 56 × 56 × 16 | Depth-wise and Point-wise conv | |
Conv3_x | 56 × 56 × 32 | Depth-wise and Point-wise conv | |
Conv4_x | 56 × 56 × 64 | Depth-wise and Point-wise conv | |
CBAM | 56 × 56 × 64 | Attention Block | |
Conv5_x | 56 × 56 × 128 | Depth-wise and Point-wise conv | |
1 × 1 × 128 | Average pool, Dropout = 0.3 | ||
Final output | 7-d fc |
Hyper-Parameters | Description |
---|---|
Batch size | 32 |
Optimiser | Adam |
Learning rate | 0.0001 |
Epoch | 150 |
Weight decay | 0.01 |
Loss function | Cross-Entropy Loss |
Channel Numbers | GFLOPs | Number of Parameters | Model Size | Average Accuracy (%) |
---|---|---|---|---|
{8, 16, 32, 64} | 0.06 | 14,473 | 0.11 MB | 94.63 |
{16, 32, 64, 128} | 0.18 | 48,041 | 0.24 MB | 96.71 |
{32, 64, 128, 256} | 0.60 | 172,777 | 0.72 MB | 96.98 |
Number of Layers in Each Residual Block | GFLOPs | Number of Parameters | Model Size | Average Accuracy (%) |
---|---|---|---|---|
(1, 2, 2, 1) | 0.10 | 23,161 | 0.12 MB | 96.13 |
(2, 2, 2, 2) | 0.16 | 41,673 | 0.21 MB | 96.19 |
(2, 3, 3, 2) | 0.18 | 48,041 | 0.24 MB | 96.71 |
(3, 4, 4, 3) | 0.26 | 72,921 | 0.36 MB | 95.54 |
Attention Module Position | GFLOPs | Number of Parameters | Model Size | Average Accuracy (%) |
---|---|---|---|---|
No CBAM block | 0.18 | 47,431 | 0.23 MB | 96.17 |
CBAM between 1st and 2nd residual layer | 0.18 | 47,561 | 0.24 MB | 96.00 |
CBAM between 2nd and 3rd residual layer | 0.18 | 47,657 | 0.24 MB | 97.2 |
CBAM between 3rd and 4th residual layer | 0.18 | 48,041 | 0.24 MB | 96.71 |
CBAM between all residual layer | 0.18 | 48,397 | 0.25 MB | 95.28 |
Class Name | False Positive Rate (FPR) | False Negative Rate (FNR) |
---|---|---|
Normal | 0.0058 | 0.0029 |
Iron | 0.0059 | 0.0492 |
Zinc | 0.0062 | 0.0704 |
Phosphorus | 0.0042 | 0.0334 |
Nitrogen | 0.0022 | 0.0161 |
MSV | 0.0023 | 0.0475 |
MLN | 0.0121 | 0.0178 |
Model | GFLOPs | Number of Parameters | Model Size | Average Accuracy (%) |
---|---|---|---|---|
SqueezeNet | 0.26 | 726,087 | 2.97 MB | 96.17 |
EfficientNetB0 | 0.39 | 4,668,035 | 18.22 MB | 96.64 |
Our Model | 0.18 | 48,041 | 0.24 MB | 96.71 |
Reference | Crop Name | Disease | Nutrition Deficiency | Dataset | Algorithm | Accuracy | Comparison with This Research Based on Performance | ||
---|---|---|---|---|---|---|---|---|---|
Number | Classes | Pros | Cons | ||||||
Wang et al. [24] | Rice | - | ✓ | 1500 | 3 | CNN with Reinforcement learning | 97% | Marginally higher accuracy | Computationally intensive |
Ibrahim et al. [22] | Palm | - | ✓ | 350 | 6 | CNN | Mean 94.29% | - | Low accuracy and complex architecture |
Bera et al. [21] | Banana Coffee Potato | ✓ | ✓ | 3076 1000 2879 | 8 9 7 | Graph Based ResNet-50 | 90%, 90.54%, 96.18% | - | Low accuracy and complex architecture |
Parnal and Alvi [23] | Black gram | - | ✓ | 3000 | 5 | ResNet-50 and VGG19 | 90.49% | - | Low accuracy and complex architecture |
Leena and Saju [8] | Corn | - | ✓ | 100 | 4 | Optimised SVM | 90% on test images | Lightweight model | Low accuracy |
Chen et al. [58] | Corn | ✓ | - | 466 | 8 | Mobile-DANet (with attention) | 95.86% of local data | - | Low accuracy with higher parameter |
Ramos-Ospina et al. [25] | Corn | - | ✓ | 2433 | 3 | CNN | 96.1% | - | Lower accuracy and computationally intensive |
Liu et al. [59] | Corn | ✓ | - | 4188 | 4 | Multi-scale ResNet | 97.45% | Slightly higher accuracy | Computationally complex |
This Research | Corn | ✓ | ✓ | 4486 | 7 | Lightweight ResNet framework with DSC and CBAM | Mean 96.71% | - | - |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Timilsina, S.; Sharma, S.; Gnawali, S.; Sato, K.; Okada, Y.; Watanabe, S.; Kondo, S. CNDD-Net: A Lightweight Attention-Based Convolutional Neural Network for Classifying Corn Nutritional Deficiencies and Leaf Diseases. Electronics 2025, 14, 1482. https://doi.org/10.3390/electronics14071482
Timilsina S, Sharma S, Gnawali S, Sato K, Okada Y, Watanabe S, Kondo S. CNDD-Net: A Lightweight Attention-Based Convolutional Neural Network for Classifying Corn Nutritional Deficiencies and Leaf Diseases. Electronics. 2025; 14(7):1482. https://doi.org/10.3390/electronics14071482
Chicago/Turabian StyleTimilsina, Suresh, Sandhya Sharma, Samir Gnawali, Kazuhiko Sato, Yoshifumi Okada, Shinya Watanabe, and Satoshi Kondo. 2025. "CNDD-Net: A Lightweight Attention-Based Convolutional Neural Network for Classifying Corn Nutritional Deficiencies and Leaf Diseases" Electronics 14, no. 7: 1482. https://doi.org/10.3390/electronics14071482
APA StyleTimilsina, S., Sharma, S., Gnawali, S., Sato, K., Okada, Y., Watanabe, S., & Kondo, S. (2025). CNDD-Net: A Lightweight Attention-Based Convolutional Neural Network for Classifying Corn Nutritional Deficiencies and Leaf Diseases. Electronics, 14(7), 1482. https://doi.org/10.3390/electronics14071482