EUVIMEDEUVIMED entdecken
Crossref

Benchmarking SE-FPN Enhanced Deep CNN Backbones for Early Detection of Mango Leaf Diseases

Muhammad Suleman Memon, Mumtaz Qabulio, Nimra Memon, Muhammad Saleem Jamali

KIET Journal of Computing and Information Sciences · 01.02.2026 · 8 (2)

DOI: 10.51153/kjcis.v82.283 ↗

Kurz gefasst

Fachartikel in der Zeitschrift KIET Journal of Computing and Information Sciences, online erschienen am 01.02.2026. Verfasst von Muhammad Suleman Memon und 3 weiteren Autor:innen. Die Arbeit nennt 20 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Mango leaves are susceptible to different diseases which causes a huge impact on the quality and quantity of the fruit. This shows how crucial it is to have reliable and automated disease detection systems to help precision agriculture. In this research, we present a novel deep learning framework that combines a Squeeze-and-Excitation (SE) enhanced convolutional backbone with a Feature Pyramid Network (FPN) directed dense fusion classification head for the recognition of multi-class mango leaf diseases. The SE mechanism dynamically readjusts channel-wise feature responses to highlight distinguishing lesion traits, whereas the dense multi-branch fusion head consolidates high-level representations to enhance feature diversity and resilience. The model is trained in two steps: first, it warms up with a frozen backbone, and then it is fully fine-tuned to improve learning of features that are specific to the domain. A lot of tests done on a carefully chosen mango leaf dataset show that the proposed architecture works well with many backbone networks. For example, DenseNet121 and InceptionV3 have validation accuracies of up to 99.75% and low validation loss. A comparative analysis shows that the proposed SE–FPN framework is better than traditional CNN-based methods when it comes to accuracy, generalization, and stability. Furthermore, evaluations utilizing confusion matrices, ROC and precision-recall curves, and Grad-CAM visualizations validate the model's strong discriminative capacity and interpretability. Overall, the proposed strategy is a good, scalable, and easy-to-understand way to use precision farming to find intelligent mango leaf disease.

Zitieren

APA

Memon, M. S., Qabulio, M., Memon, N., & Jamali, M. S. (2026). Benchmarking SE-FPN Enhanced Deep CNN Backbones for Early Detection of Mango Leaf Diseases. KIET Journal of Computing and Information Sciences, 8(2). https://doi.org/10.51153/kjcis.v82.283

Vancouver

Memon MS, Qabulio M, Memon N, Jamali MS. Benchmarking SE-FPN Enhanced Deep CNN Backbones for Early Detection of Mango Leaf Diseases. KIET Journal of Computing and Information Sciences. 2026;8(2). doi:10.51153/kjcis.v82.283

BibTeX
@article{Memon2026,
  author = {Memon, Muhammad Suleman and Qabulio, Mumtaz and Memon, Nimra and Jamali, Muhammad Saleem},
  title = {Benchmarking SE-FPN Enhanced Deep CNN Backbones for Early Detection of Mango Leaf Diseases},
  journal = {KIET Journal of Computing and Information Sciences},
  year = {2026},
  volume = {8},
  number = {2},
  doi = {10.51153/kjcis.v82.283}
}
Kostenlos starten →