Combining RESNET50 and SAM in Detecting Rubber Tree Leaf Diseases
PDF

Keywords

Rubber Tree
Leaf Disease
RESNET50 CNN
Segment Anything Model
Image Classification
Image Segmentation

Abstract

This work proposes a deep learning-based approach to detect and identify rubber tree leaf diseases by combining Segment Anything Model (SAM) and a fine-tuned RESNET50 convolutional neural network. Rubber tree leaf images were collected from a plantation in Sta. Maria, Laguna, Philippines within a natural field condition to capture variations in lighting, background, and leaf orientation. SAM was applied to isolate the leaf regions from background elements, creating a curated dataset of 423 validated images used for model training and evaluation. The classification model utilized transfer learning with a pretrained RESNET50 architecture and was evaluated in terms of accuracy, precision, recall, and F1-score. Study results show that the proposed RESNET50 model achieved an overall classification accuracy of 72%, demonstrating promising detection performance some of the image classes. The findings indicate that combining segmentation and deep convolutional classification can support automated and scalable rubber leaf disease identification in real plantation environments.

PDF

References

[1] Department of Trade and Industry. 2023. Philippine rubber

industry roadmap 2023–2028. Republic of the Philippines.

https://dtiwebfiles.s3.ap-southeast-1.amazonaws.com/e

library/Other+Publications/Philippine+Rubber+Roadmap+2023+2

028.pdf

[2] Philippine Statistics Authority. (2023). Rubber (cup lump)

production. April to June 2023. https://psa.gov.ph/major-non

food-industrial-crops/rubber

[3] Philippine Council for Agriculture, Aquatic and Natural Resources

Research and Development (PCAARRD). (2024). Rubber industry

profile. Industry Strategic Science and Technology Plans (ISPs)

Platform.

https://ispweb.pcaarrd.dost.gov.ph/isp

commodities/rubber/

[4] Solpot, T. C., Borja, B. T., Prado, M. M., Abubakar, J. V., &

Cabasan, M. T. N. (2024). Leaf diseases of Hevea brasiliensis

M ̈ull. Arg. in major rubber growing areas of Cotabato,

Philippines.

Journal of Rubber Research, 27, 11–20.

https://doi.org/10.1007/s42464-023-00230-6

[5] Mozter, P. (2023, January 25). Rubber trees has been ravaged by

an infectious disease that could affect 50,000 hectares of

plantation in Basilan, Philippines. Nature World News.

https://www.natureworldnews.com/articles/55199/20230125/

rubber-trees-ravaged-infectious-disease-affect-50-000-hectares

plantation.htm

[6] Aliya, S. S. S., Syd Ali, N., Mahyudin, M. M., Wong, M. Y., &

Yusop, M. R. (2022). Emerging and existing major leaf diseases

of Hevea brasiliensis in Malaysia. Journal of Current Opinion in

https://www.researchgate.net/publication/359576642_Emerging_a

nd_Existing_Major_Leaf_Diseases_of_Hevea_brasiliensis_in_Ma

laysia

[7] Pornsuriya, C., Thaochan, N., Chairin, T., & Sunpapao, A. (2023).

Morphological and phylogenetic evidences reveal Lasiodiplodia

chonburiensis and L. theobromae associated with leaf blight in

Hevea brasiliensis in southern Thailand. Diversity, 15(9), 961.

https://doi.org/10.3390/d15090961

[8] Thaochan, N., Pornsuriya, C., Chairin, T., & Community, P.

(2022). Morphological and molecular characterization of

Calonectria foliicola associated with leaf blight on rubber tree

(Hevea brasiliensis) in Thailand. Journal of Fungi, 8(10), 986.

https://doi.org/10.3390/jof8100986

[9] Shi, P., Qiu, J., Mu, S. D., Wei, H., Lo, F. P.-W., & Yuan, W.

(2023). Generalist vision foundation models for medical imaging:

A case study of Segment Anything Model on zero-shot medical

segmentation

(arXiv:2304.12637).

org/10.48550/arXiv.2304.12637

arXiv.

https://doi.

[10] Chen, T., Zhu, L., Ding, C., Cao, R., Wang, Y., Li, Z., Sun, L.,

Mao, P., & Zang, Y. (2023). SAM fails to segment anything? —

SAM-Adapter: Adapting SAM in underperformed scenes:

Camouflage, shadow, medical image segmentation, and more

(arXiv:2304.09148).

arXiv.

https://doi.org/10.48550/arXiv.2304.09148

[11] Moupojou, E., Retraint, F., Tapamo, H., Nkenlifack, M., Kacfah,

C., & Tagne, A. (2024). Segment Anything Model & Fully

Convolutional Data Description for Plant Multi-disease Detection

on Field Images. IEEE Access. https://doi.org/10.1109/

ACCESS.2024.3433495

[12] Krishna, M. S., Machado, P., Otuka, R. I., Yahaya, S. W., dos

Santos, F. N., & Ihianle, I. K. (2025). Plant leaf disease detection

using deep learning: A multi-dataset approach. J -

Multidisciplinary

Scientific

https://doi.org/10.3390/j8010004

Journal,

8(1),

4.

[13] Singh, B., Sharma, R. P., & Dey, S. (2025). Interpretable plant leaf

disease

detection

(arXiv:2512.17864).

48550/arXiv.2512.17864

using

attention-enhanced

arXiv.

CNN

https://doi.org/10.

[14] Aboelenin, S., Elbasheer, F. A., Eltoukhy, M. M., El-Hady, W.

M., & Hosny, K. M. (2025). A hybrid framework for plant leaf

disease detection and classification using convolutional neural

networks and vision transformer. Complex & Intelligent Systems,

11, Article 142. https://doi.org/10.1007/s40747-024-01764-x

[15] Zeng, T., Li, C., Zhang, B., Wang, R., Fu, W., Wang, J., & Zhang,

X. (2022). Rubber leaf disease recognition based on improved

deep convolutional neural networks with a cross-scale attention

mechanism. Frontiers in Plant Science, 13, 829479. https://doi.

org/10.3389/fpls.2022.829479

[16] Lertkrai, P., Chanakot, B., & Kaewboonma, N. (2025).

Deployment of a deep learning model for the automated diagnosis

of Thai rubber leaf diseases via the LINE platform. International

Journal of Computer Theory and Engineering, 17(3), 126–133.

https: //doi.org/10.7763/IJCTE.2025.V17.1375

[17] Kirillov, A., Mintun, E., Ravi, N., Mao, H., Rolland, C.,

Gustafson, L., Xiao, T., Whitehead, S., Berg, A. C., Lo, W.-Y.,

Dollár, P., & Girshick, R. (2023). Segment anything

(arXiv:2304.02643).

arXiv.

https://doi.org/10.48550/arXiv.2304.02643

[18] He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual

learning for image recognition. In Proceedings of the IEEE

conference on computer vision and pattern recognition (pp. 770

778).

Creative Commons License

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Copyright (c) 2026 Daniel Fesalbon, Arvin (Author)