Agricultural Disclaimer: This tool is intended for research and informational purposes only. Final disease identification and treatment protocols should be verified by a qualified agronomist or plant pathologist before application.
MustardSense provides an integrated, dual-stream deep learning paradigm designed to bridge the gap between post-harvest mustard quality assessment and real-time field diagnostics.
Our platform utilizes a novel dual-stream explainable deep network designed to handle fine-grained mustard powder quality grading and complex aphid infestation categories simultaneously. This architecture provides a unified solution for both crop health monitoring and processing quality.
We prioritize transparency using Grad-CAM++ visualization to validate the model's decision-making process. This ensures that users can see exactly which leaf textures or aphid-dense regions the AI is focusing on for its assessment.
By integrating attention-driven feature refinement, the system learns to emphasize critical spatial patterns, such as early fungal spread and aphid clusters. This allows for more robust visual understanding than conventional deep learning models.
Our network employs Squeeze-and-Excitation (SE) modules to adaptively recalibrate channel responses. This technical optimization reweights feature channels based on global context, ensuring that visually dominant disease and pest patterns receive the highest priority.
The model incorporates Ghost Convolutions to enhance computational efficiency while maintaining high accuracy. This makes MustardSense a scalable, low-cost, and reliable digital tool suitable for real-world agricultural environments.
MustardSense classifies symptoms into granular severity levels, including five distinct scores for Powdery Mildew and five population density ranges for Aphid infestations. This fine-grained classification empowers farmers to make smarter, data-driven management decisions..
...........
Meet the researchers and experts leading the MustardSense initiative at ACAI (Amity University) & ICAR-National Institute of Biotic Stress Management.
Project Research Scientist, Amity Centre for Artificial Intelligence (ACAI), Amity University, Noida.
Joint Director, ICAR-National Institute of Biotic Stress Management, Chhattisgarh.
Assistant Professor, Amity Centre for Artificial Intelligence (ACAI), Amity University, Noida.
Professor, Amity Centre for Artificial Intelligence (ACAI), Amity University, Noida.
Scientist ( Environmental Science) School of Crop Health Biology Research ICAR- NIBSM.
Have questions or need assistance? Fill out the form below and our team will get back to you ASAP.
