EUVIMEDEUVIMED entdecken
Crossref

COMPARATIVE ANALYSIS OF DEEP LEARNING ARCHITECTURES FOR COTTON PLANT DISEASE DETECTION USING GRAD-CAM VISUALIZATION

Poonam U. Gadgil, L. B. Patle, Kunal D. Gaikwad

Research Hub · 01.01.2026 · 07 (02) · 320-327

DOI: 10.63665/rh.v7i2.51 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Research Hub, online erschienen am 01.01.2026. Verfasst von Poonam U. Gadgil und 2 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Early and accurate identification of plant diseases is critical for improving crop yield and ensuring sustainable agricultural practices. Cotton, being a major commercial crop, is highly vulnerable to leaf diseases that significantly impact productivity. Deep learning has recently emerged as a reliable solution for automated plant disease diagnosis using imagebased approaches [1], [2]. This study presents a comparative analysis of two deep learning architectures—DenseNet121 and MobileNetV2—for automated cotton plant disease classification using leaf images. The experiments were conducted on the publicly available Cotton Plant Disease Dataset by Dhamodharan R., consisting of approximately 4,800 images spanning five disease categories and one healthy class. Transfer learning was employed with ImageNet pre-trained weights, and models were trained using a batch size of 32, learning rate of 1×10⁻⁴, and GPU acceleration. Performance evaluation was carried out using accuracy, macro precision, macro recall, and macro F1-score metrics. DenseNet121 achieved superior performance with an accuracy of 98.32%, outperforming MobileNetV2 while maintaining robust generalization. Additionally, Gradient-weighted Class Activation Mapping (Grad-CAM) was utilized to provide visual interpretability by highlighting diseaseaffected regions on cotton leaves. The results demonstrate that DenseNet121 offers an effective and explainable solution for cotton disease detection, while MobileNetV2 provides a lightweight alternative suitable for resource-constrained environments

Zitieren

APA

Gadgil, P. U., Patle, L. B., & Gaikwad, K. D. (2026). COMPARATIVE ANALYSIS OF DEEP LEARNING ARCHITECTURES FOR COTTON PLANT DISEASE DETECTION USING GRAD-CAM VISUALIZATION. Research Hub, 07(02), 320-327. https://doi.org/10.63665/rh.v7i2.51

Vancouver

Gadgil PU, Patle LB, Gaikwad KD. COMPARATIVE ANALYSIS OF DEEP LEARNING ARCHITECTURES FOR COTTON PLANT DISEASE DETECTION USING GRAD-CAM VISUALIZATION. Research Hub. 2026;07(02):320-327. doi:10.63665/rh.v7i2.51

BibTeX
@article{Gadgil2026,
  author = {Gadgil, Poonam U. and Patle, L. B. and Gaikwad, Kunal D.},
  title = {COMPARATIVE ANALYSIS OF DEEP LEARNING ARCHITECTURES FOR COTTON PLANT DISEASE DETECTION USING GRAD-CAM VISUALIZATION},
  journal = {Research Hub},
  year = {2026},
  volume = {07},
  number = {02},
  pages = {320--327},
  doi = {10.63665/rh.v7i2.51}
}
Kostenlos starten →