Effectiveness of different backbones for the DeepLabV3+ network applied to semantic segmentation of agricultural crops with limited image data
Abstract
This study investigates the performance of ResNet-50, ConvNeXt, and Vision Transformer (ViT) for the DeepLabV3+ network, applied to the semantic segmentation of images of corn crops with weeds, using a reduced set of 58 images. To mitigate data scarcity, techniques such as data augmentation and the combination of two loss functions were implemented. The models were evaluated in terms of accuracy and mean Intersection over Union (mIoU). ResNet-50 achieved the highest overall mIoU on the test set (0.7255), but ViT proved superior in weed identification, highlighting its potential to capture minute details. This article contributes to optimizing weed identification and reducing the use of pesticides.
References
Ai Studio Dataset. Disponível em: [link]
Carnevalli, S. S. e S. (2020). “Estudo de Deep Learning para o reconhecimento de ervas invasoras na cultura do milho a partir de imagens de VANTs”. 2020. 35f. Trabalho de Conclusão de Curso, UNESP, Tupã.
Ford, J., Sadgrove, E. e Paul, D. (2025). "Joint plant-spraypoint detector with ConvNeXt modules and HistMatch normalization". Precision Agriculture, v. 26, art. 24, jan. 2025. DOI: 10.1007/s11119-024-10208-y
Garcia-Garcia, A., Orts-Escolano, S., Oprea, S. O., Villena-Martinez, V. e GarciaRodriguez, J. (2017). "A Review on Deep Learning Techniques Applied to Semantic Segmentation". ArXiv preprint arXiv:1704.06857.
Huang, X., Xu, D., Chen, Y., Zhang, Q., Feng, P., Ma, Y., Dong, Q. e Yu, F. (2025). EConv-ViT: A strongly generalized apple leaf disease classification model based on the fusion of ConvNeXt and Transformer, “Information Processing in Agriculture”, 2025, ISSN 2214-3173.
Jiang, K., Afzaal, U. e Lee, J. (2022). "Transformer-Based Weed Segmentation for Grass Management". Sensors, 2023, v. 23, n. 1, art. 65, dez. 2022. DOI: 10.3390/s23010065
Lecun, Y., Bengio, Y. e Hinton, G. (2015). “Deep Learning”. Nature, v. 521. f. 436-444.
Pereira, G. F. D. (2023). “Investigação de algoritmos de aprendizado profundo para a segmentação semântica de plantas daninhas e culturas agrícolas usando imagens de VANTs”. 2023. Trabalho de Conclusão de Curso, UNESP, Itapeva.
Rezaei, M., Diepeveen, D., Laga, H., Jones, M. G. K. e Sohel, F. (2024). “Plant disease recognition in a low data scenario using few-shot learning”, Computers and Electronics in Agriculture, v. 219, 2024, ISSN 0168-1699.
Taha, H., El-Habrouk, H., Bekheet, W., El-Naghi, S. e Torki, M. (2025). “Pixel-level pavement crack segmentation using UAV remote sensing images based on the ConvNeXt-UPerNet”, Alexandria Engineering Journal, V. 124, 2025, p. 147-169, ISSN 1110-0168.
Thapa, R., Zhang, K., Snavely, N., Belongie, S., Khan, A. (2020) “The Plant Pathology Challenge 2020 data set to classify foliar disease of apples”. Appl Plant Sci 2020; 8:e11390.
Wu, Q., Ma, X., Liu, H., Bi, C., Yu, H., Liang, M., Zhang, J., Li, Q., Tang, Y. e Ye, G. (2023). “A classification method for soybean leaf diseases based on an improved ConvNeXt model”. Sci Rep 13, 19141 (2023).
Yang, Q., Duan, S., Wang, L. (2022) “Efficient identification of apple leaf diseases in the wild using convolutional neural networks. Agronomy 2022; 12:2784.
Yun, S., Han, D., Oh, S. J., Chun, S., Choe, J., Yoo, Y. (2019). “CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features” IEEE/CVF International Conference on Computer Vision (ICCV), 2019, Seul, Coreia do Sul.
Zhu, H., Chen, B. e Yang, C. (2023). “Understanding Why ViT Trains Badly on Small Datasets: An Intuitive Perspective”, ArXiv: 2302.03751.
Zhuang, L., Mao, H., Wu, C.-Y., Feichtenhofer, C., Darrell, T., e Xie, S. (2022) “A ConvNet for the 2020s”. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (CVPR). 2022. p. 11976-11986.
Carnevalli, S. S. e S. (2020). “Estudo de Deep Learning para o reconhecimento de ervas invasoras na cultura do milho a partir de imagens de VANTs”. 2020. 35f. Trabalho de Conclusão de Curso, UNESP, Tupã.
Ford, J., Sadgrove, E. e Paul, D. (2025). "Joint plant-spraypoint detector with ConvNeXt modules and HistMatch normalization". Precision Agriculture, v. 26, art. 24, jan. 2025. DOI: 10.1007/s11119-024-10208-y
Garcia-Garcia, A., Orts-Escolano, S., Oprea, S. O., Villena-Martinez, V. e GarciaRodriguez, J. (2017). "A Review on Deep Learning Techniques Applied to Semantic Segmentation". ArXiv preprint arXiv:1704.06857.
Huang, X., Xu, D., Chen, Y., Zhang, Q., Feng, P., Ma, Y., Dong, Q. e Yu, F. (2025). EConv-ViT: A strongly generalized apple leaf disease classification model based on the fusion of ConvNeXt and Transformer, “Information Processing in Agriculture”, 2025, ISSN 2214-3173.
Jiang, K., Afzaal, U. e Lee, J. (2022). "Transformer-Based Weed Segmentation for Grass Management". Sensors, 2023, v. 23, n. 1, art. 65, dez. 2022. DOI: 10.3390/s23010065
Lecun, Y., Bengio, Y. e Hinton, G. (2015). “Deep Learning”. Nature, v. 521. f. 436-444.
Pereira, G. F. D. (2023). “Investigação de algoritmos de aprendizado profundo para a segmentação semântica de plantas daninhas e culturas agrícolas usando imagens de VANTs”. 2023. Trabalho de Conclusão de Curso, UNESP, Itapeva.
Rezaei, M., Diepeveen, D., Laga, H., Jones, M. G. K. e Sohel, F. (2024). “Plant disease recognition in a low data scenario using few-shot learning”, Computers and Electronics in Agriculture, v. 219, 2024, ISSN 0168-1699.
Taha, H., El-Habrouk, H., Bekheet, W., El-Naghi, S. e Torki, M. (2025). “Pixel-level pavement crack segmentation using UAV remote sensing images based on the ConvNeXt-UPerNet”, Alexandria Engineering Journal, V. 124, 2025, p. 147-169, ISSN 1110-0168.
Thapa, R., Zhang, K., Snavely, N., Belongie, S., Khan, A. (2020) “The Plant Pathology Challenge 2020 data set to classify foliar disease of apples”. Appl Plant Sci 2020; 8:e11390.
Wu, Q., Ma, X., Liu, H., Bi, C., Yu, H., Liang, M., Zhang, J., Li, Q., Tang, Y. e Ye, G. (2023). “A classification method for soybean leaf diseases based on an improved ConvNeXt model”. Sci Rep 13, 19141 (2023).
Yang, Q., Duan, S., Wang, L. (2022) “Efficient identification of apple leaf diseases in the wild using convolutional neural networks. Agronomy 2022; 12:2784.
Yun, S., Han, D., Oh, S. J., Chun, S., Choe, J., Yoo, Y. (2019). “CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features” IEEE/CVF International Conference on Computer Vision (ICCV), 2019, Seul, Coreia do Sul.
Zhu, H., Chen, B. e Yang, C. (2023). “Understanding Why ViT Trains Badly on Small Datasets: An Intuitive Perspective”, ArXiv: 2302.03751.
Zhuang, L., Mao, H., Wu, C.-Y., Feichtenhofer, C., Darrell, T., e Xie, S. (2022) “A ConvNet for the 2020s”. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (CVPR). 2022. p. 11976-11986.
Published
2025-09-29
How to Cite
PEREIRA, Gabriel F. D.; COLETTA, Luiz F. S.; ROSSI, André L. D..
Effectiveness of different backbones for the DeepLabV3+ network applied to semantic segmentation of agricultural crops with limited image data. In: NATIONAL MEETING ON ARTIFICIAL AND COMPUTATIONAL INTELLIGENCE (ENIAC), 22. , 2025, Fortaleza/CE.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2025
.
p. 604-615.
ISSN 2763-9061.
DOI: https://doi.org/10.5753/eniac.2025.13943.
