DWSC-SE: A Lightweight Depthwise Separable Convolution enhanced using Squeeze and Excitation Blocks for Rice Leaf Disease Detection
Dandao Xuebao/Journal of Ballistics · 29.05.2026 · 38 (2) · 43-50
DOI: 10.52783/dxjb.v38.309 ↗Kurz gefasst
Fachartikel in der Zeitschrift Dandao Xuebao/Journal of Ballistics, online erschienen am 29.05.2026. Verfasst von Baranidharan B.
Abstract (Originalsprache)
Rice is eaten by over half the world's population because of its nutritional value. However, rice leaf diseases, affect its yield quantity. Rice yield loss can be mitigated more effectively if detected early. This research paper explores the effectiveness of lightweight deep learning-based models against heavyweight models for automated detection of rice leaf diseases. A publicly available rice leaf disease dataset was used, comprising labeled images of four varieties of diseased leaves. Pretrained models like ResNet152 and DenseNet201 models are compared with Custom CNN, Depth Wise Separable Convolution (DWSC), and Depth Wise Separable Convolution – Squeeze & Excitation (DWSC-SE) models. ResNet152 and DenseNet201 are considered heavyweight models due to their large number of parameters, whereas the other three are lightweight. It is observed that separable convolutional models perform better than other CNN models with a very small number of parameters. Particularly, DWSC-SE is showing better results than all other CNN models. The proposed model demonstrates high accuracy in classifying the leaf images and outperforms traditional deep learning methods
Zitieren
APA
B, B. (2026). DWSC-SE: A Lightweight Depthwise Separable Convolution enhanced using Squeeze and Excitation Blocks for Rice Leaf Disease Detection. Dandao Xuebao/Journal of Ballistics, 38(2), 43-50. https://doi.org/10.52783/dxjb.v38.309
Vancouver
B B. DWSC-SE: A Lightweight Depthwise Separable Convolution enhanced using Squeeze and Excitation Blocks for Rice Leaf Disease Detection. Dandao Xuebao/Journal of Ballistics. 2026;38(2):43-50. doi:10.52783/dxjb.v38.309
BibTeX
@article{B2026,
author = {B, Baranidharan},
title = {DWSC-SE: A Lightweight Depthwise Separable Convolution enhanced using Squeeze and Excitation Blocks for Rice Leaf Disease Detection},
journal = {Dandao Xuebao/Journal of Ballistics},
year = {2026},
volume = {38},
number = {2},
pages = {43--50},
doi = {10.52783/dxjb.v38.309}
}
EUVIMED