JassidNet as a quantization-aware lightweight phenotyping framework for high-throughput cotton jassid (Amrasca biguttula) detection and counting toward objective resistance screening
6 affiliations
- 1Bio-Sensing, Automation, and Intelligence Laboratory, Department of Agricultural and Biological Engineering, Institute of Food and Agricultural Sciences, University of Florida, Gainesville, 32611, FL, USA
- 2Department of Crop and Soil Sciences, College of Agriculture and Environmental Sciences, University of Georgia, Tifton, 31793, GA, USA
- 3Department of Computer Science, Wake Forest University, 1834 Wake Forest Road, Winston-Salem, 27109, NC, USA
- 4Institute of Plant Breeding, Genetics and Genomics, University of Georgia-Tifton Campus, Tifton, 31793, GA, USA
- 5Cotton Incorporated, Cary, 27513, NC, USA
- 6Department of Entomology, University of Georgia, Tifton, 31793, GA, USA
Abstract
Cotton jassid ( Amrasca biguttula ) poses a growing threat to cotton production worldwide, which creates an urgent demand for scalable and field-deployable tools that support population monitoring and resistance evaluation under natural conditions. This study proposes JassidNet, the first high-throughput computer vision-based deep learning neural network for automated cotton jassid detection and counting in real-world field environments. JassidNet integrates a lightweight detection model (SJIMNet) with standardized preprocessing, optional super-resolution enhancement, and flexible segmentation strategies, including detector-guided and SAM3-based text-driven zero-shot segmentation, to ensure robustness under heterogeneous acquisition conditions. To support model development and evaluation, a high-quality Cotton Jassid Recognition (CJR) dataset was constructed using a semi-supervised pseudo-label iteration strategy, substantially reducing manual annotation effort while maintaining label reliability. SJIMNet incorporates SPDConv, a self-customized GIAM-Block, and quantization-aware training, achieving comparable or superior detection performance relative to the YOLOv11m baseline while reducing computational cost by 20.6%, parameter count by 26.1%, and model size by 61.1%. The quantized model further compresses storage requirements to 15,388 KB while largely preserving counting-related performance, enabling efficient deployment on mobile and edge devices. The quantized model achieved 116.188 ms latency and 8.608 FPS on a CPU-only edge-computing platform, further supporting its practical deployment under resource-constrained field conditions. Explainable AI-based feature response analysis shows that, despite its lightweight design, JassidNet more effectively concentrates attention on pest-relevant regions. Extensive evaluations on both internal and independent external datasets demonstrate strong generalization capability, yielding an overall Mean Absolute Error (MAE) of 1.7, a Root Mean Square Error (RMSE) of 2.21 jassids per sampling unit, and a Pearson correlation coefficient of 0.974. Field validations further reveal a statistically consistent association between image-derived jassid counts across diverse cotton genotypes, supporting genotype-level tolerance discrimination under natural infestation pressure. Overall, JassidNet provides a practical and interpretable solution for high-throughput jassid monitoring and resistance screening, establishing a methodological foundation for intelligent plant protection and digital agriculture. Both the code and dataset can be accessed at: https://github.com/UGA-BSAIL/JassidNet .
Related
- Closing the phenotyping gap: non-invasive electromagnetic induction enables high-throughput screening of crop root function in the field
- Model-assisted high-throughput phenotyping of photosynthetic acclimation
- Non-destructive estimation of individual buckwheat plant biomass using high-throughput phenotyping-derived canopy traits
- Physiology-informed high-throughput phenotyping of grain moisture dynamics provides enhanced insights into rice grain weight formation
- A high-throughput UAV millimetric imaging pipeline for plant density phenotyping reveals high heritability in wheat-pea sole crops and intercrops.
BibTeX
@article{wang2026jassidnet,
title = {{JassidNet as a quantization-aware lightweight phenotyping framework for high-throughput cotton jassid (Amrasca biguttula) detection and counting toward objective resistance screening}},
author = {Rui-Feng Wang and Kangning Cui and Iago Beffart Schardong and Matthew C. Bauer and Rama Vamsi Somala and Mingrui Xu and Dalton West and Don C. Jones and Sally V. Taylor and Phillip M. Roberts and Changying Li and Peng Wah Chee},
journal = {Industrial Crops and Products},
year = {2026},
month = jun,
doi = {10.1016/j.indcrop.2026.123716},
url = {https://doi.org/10.1016/j.indcrop.2026.123716},
note = {CC BY-NC-ND},
}
RIS
TY - JOUR
TI - JassidNet as a quantization-aware lightweight phenotyping framework for high-throughput cotton jassid (Amrasca biguttula) detection and counting toward objective resistance screening
AU - Rui-Feng Wang
AU - Kangning Cui
AU - Iago Beffart Schardong
AU - Matthew C. Bauer
AU - Rama Vamsi Somala
AU - Mingrui Xu
AU - Dalton West
AU - Don C. Jones
AU - Sally V. Taylor
AU - Phillip M. Roberts
AU - Changying Li
AU - Peng Wah Chee
PY - 2026
DA - 2026/06/22
JO - Industrial Crops and Products
DO - 10.1016/j.indcrop.2026.123716
UR - https://doi.org/10.1016/j.indcrop.2026.123716
AB - Cotton jassid ( Amrasca biguttula ) poses a growing threat to cotton production worldwide, which creates an urgent demand for scalable and field-deployable tools that support population monitoring and resistance evaluation under natural conditions. This study proposes JassidNet, the first high-throughput computer vision-based deep learning neural network for automated cotton jassid detection and counting in real-world field environments. JassidNet integrates a lightweight detection model (SJIMNet) with standardized preprocessing, optional super-resolution enhancement, and flexible segmentation strategies, including detector-guided and SAM3-based text-driven zero-shot segmentation, to ensure robustness under heterogeneous acquisition conditions. To support model development and evaluation, a high-quality Cotton Jassid Recognition (CJR) dataset was constructed using a semi-supervised pseudo-label iteration strategy, substantially reducing manual annotation effort while maintaining label reliability. SJIMNet incorporates SPDConv, a self-customized GIAM-Block, and quantization-aware training, achieving comparable or superior detection performance relative to the YOLOv11m baseline while reducing computational cost by 20.6%, parameter count by 26.1%, and model size by 61.1%. The quantized model further compresses storage requirements to 15,388 KB while largely preserving counting-related performance, enabling efficient deployment on mobile and edge devices. The quantized model achieved 116.188 ms latency and 8.608 FPS on a CPU-only edge-computing platform, further supporting its practical deployment under resource-constrained field conditions. Explainable AI-based feature response analysis shows that, despite its lightweight design, JassidNet more effectively concentrates attention on pest-relevant regions. Extensive evaluations on both internal and independent external datasets demonstrate strong generalization capability, yielding an overall Mean Absolute Error (MAE) of 1.7, a Root Mean Square Error (RMSE) of 2.21 jassids per sampling unit, and a Pearson correlation coefficient of 0.974. Field validations further reveal a statistically consistent association between image-derived jassid counts across diverse cotton genotypes, supporting genotype-level tolerance discrimination under natural infestation pressure. Overall, JassidNet provides a practical and interpretable solution for high-throughput jassid monitoring and resistance screening, establishing a methodological foundation for intelligent plant protection and digital agriculture. Both the code and dataset can be accessed at: https://github.com/UGA-BSAIL/JassidNet .
SN - 0926-6690
SN - 1872-633X
ER -