Latest News

PhD student Toqi Tahamid Sarker has been awarded the Doctoral Research Fellowship.

Two papers have been accepted at ICCV 2025. WeedSense, a multi-task architecture that simultaneously performs weed segmentation, height estimation, and growth stage classification, was accepted at the 10th Computer Vision in Plant Phenotyping and Agriculture workshop. The second paper, GasTwinFormer, introduces a hybrid vision transformer for real-time livestock methane monitoring and has been selected for oral presentation and as a best paper nominee at the SEA: Sustainability with Earth Observation & AI workshop.

We are thrilled to announce that our journal article, 'WeedSwin Hierarchical Vision Transformer with SAM-2 for Multi-Stage Weed Detection and Classification' has been published in Scientific Reports (Nature). This comprehensive research introduces extensive datasets documenting prevalent weed species across multiple growth stages, and proposes a novel WeedSwin Transformer architecture that demonstrates exceptional performance in weed detection and classification for precision agriculture. Congratulations to the team!

We are excited to announce that our review article, 'Carbon Emission Quantification of Machine Learning: A Review' has been published in IEEE Transactions on Sustainable Computing. This collaborative work with SHIELD Lab, SIU represents an important contribution to sustainable computing research. Congratulations to the entire team!

We are thrilled to announce that our journal article, 'WeedSwin Hierarchical Vision Transformer with SAM-2 for Multi-Stage Weed Detection and Classification' has been acepted by the Scientific Reports (Nature). Congratulations to the team!

We are pleased to announce that our review article, 'Carbon Emission Quantification of Machine Learning: A Review' has been acepted by the IEEE Transactions on Sustainable Computing. This is one of our collaboration work with SHIELD Lab, SIU. Congratulations to the team!

We are pleased to announce that our research proposal, 'AI-Driven Discovery of Sustainable Ionic Liquids: A Multimodal Deep Learning Approach for Enhanced Lignin and Plastic Deconstruction,' has been awarded funding by the IIN Sustaining Illinois program. The period of performance is May 16, 2025, to May 15, 2026

Our lab's latest research "Optical gas imaging and deep learning for quantifying enteric methane emissions from rumen fermentation in vitro" published in IET Image Processing explores a potential new approach to greenhouse gas monitoring in the livestock industry. By combining Optical Gas Imaging (OGI) with deep learning, we've investigated a promising alternative to traditional methane quantification methods. Our interdisciplinary collaboration between Animal Science and Computer Science shows how we can make environmental sustainability more achievable for farmers and researchers alike. While further research is needed, this initial work represents an important step in exploring more accessible approaches to livestock greenhouse gas monitoring.

Taminul Islam, a PhD student and a Research Assistant from the BASE Lab, presented our research titled "WeedVision: Multi-Stage Growth and Classification of Weeds Using DETR and RetinaNet for Precision Agriculture" at the Deep Learning and Applications special session during the 23rd International Conference on Machine Learning and Applications (ICMLA).

Taminul Islam, a PhD student and Research Assistant from the BASE Lab, presented a poster titled "WeedVision: Multi-Stage Growth and Classification of Weeds Using DETR and RetinaNet for Precision Agriculture" at the Illinois Innovation Network’s Sustainability Research Conference, held at Illinois State University.

Taminul Islam and Toqi Tahamid Sarker, PhD students and Research Assistants from the BASE Lab, presented a poster titled "Detection and Classification of Cannabis Seeds Using RetinaNet and Faster R-CNN" at Southern Illinois University’s 5th Annual Hemp Cannabis Symposium 2024.

Awarded a one-year research grant from Illinois Soybean Center for "Smart Farming Soybean: The use of AI to detect and estimate early insect infections in Soybean"

Toqi Tahamid Sarker presented our research titled "Gasformer: A Transformer-based Architecture for Segmenting Methane Emissions from Livestock in Optical Gas Imaging" at the 5th International Workshop on Agriculture-Vision: Challenges & Opportunities for Computer Vision in Agriculture, held in conjunction with IEEE/CVF CVPR 2024. The research was presented both as an oral presentation and during the poster session. The workshop brought together leading researchers in agricultural computer vision applications to discuss advances in the field.

Awarded a two-year research federal grant from USDA-NIFA for "Detecting Subacute Ruminal Acidosis using a real-time Deep Learning Algorithm"

PhD student Memari Majid successfully defended his dissertation and will graduate in Fall 2023

Granted a one-year seed research grant from ILLINOIS INNOVATION NETWORK.

Secured a two-year research federal grant from USDA-NIFA titled "Artificial Intelligence for Greener Livestock: Educational and Research."

Graduate student Subash Kharel received the Outstanding Thesis Award 2021.