Research
My group builds computer vision and machine learning methods that work outside the lab: in farm fields, from drones and satellites, and on small edge devices.
Themes
Precision agriculture and plant phenotyping
Segmenting crops, weeds and crop damage; counting plants, wheat spikes and kernels; detecting Fusarium head blight and lodging; monitoring flea beetles, wheat midge and other insects from trap images; and estimating grain yield and harvest index from UAV multispectral imagery for wheat breeding programs.
Learning with less labelled data
Unsupervised domain adaptation across crops, fields and seasons; semi-supervised segmentation guided by language; and parameter-efficient adaptation of foundation models such as the Segment Anything Model.
Soil and spectroscopy
Machine learning for laser-induced breakdown spectroscopy to classify soils and measure carbon, nitrogen and bitumen content, and mapping soil organic matter and surface texture from RGB images.
Beyond agriculture
Camouflaged object detection, EEG and multimodal signals for health, machine-learning-driven microwave filter design, acoustic monitoring for mine safety, and analysis of social media from conflict zones.
Current projects
- A Long-Term Management Plan for Herbicide-Resistant Kochia and Wild OatSaskatchewan Ministry of Agriculture, ADF Strategic Research Initiative; SaskOilSeeds, SaskWheat, RDAR, Alberta Grains, SaskBarley, Sask Pulse Growers, ADF Strategic Research Initiative2026–2031PartnerCollaborator
- Developing an AI Platform for Selecting Abiotic Stress Resilient Traits in WheatSK Wheat Development Commission and SK Ministry of Agriculture, Agriculture Development Fund2026–2030LeadPrincipal investigator
- Crop Stress Management using Multisource Data FusionNSERC, Discovery Launch Supplement and Discovery Grant2021–2027LeadPrincipal investigator
Completed projects
- Using Machine Learning to Improve Underground Mine Roof Safety EvaluationGIEMS, Collaborative Energy, Minerals and Society Research Award2025PartnerCo-investigator
- Improved Perception and Navigation for Automated Harvesting using Imaging SensorsMacDon Industries Ltd, Alliance–Mitacs Accelerate; NSERC and Mitacs, Alliance–Mitacs Accelerate2024–2025PartnerCo-applicant
- Developing Machine Learning Methods for RGB Images to Quantify Crop and Weed Populations Across Agricultural FieldsCroptimistic Technology Inc.; NSERC and Mitacs, Alliance–Mitacs Accelerate2023–2026LeadPrincipal investigator
- Generative Codecs for Smart Video EncodingMitacs, Accelerate2023–2025LeadPrincipal investigator
- Image Based Detection of Fusarium Head Blight, Lodging, and Grain Yield to Aid Phenotypic Selections in Wheat BreedingSK Wheat Development Commission, SK Ministry of Agriculture, MB Crop Alliance, Agriculture Development Fund2022–2026LeadPrincipal investigator
- Laser Induced Breakdown Spectroscopy for Efficient Germplasm Selection of Wheat Grain AttributesSaskatchewan Wheat Development Commission and Western Grains Research Foundation, Agriculture Development Fund2022–2025PartnerCollaborator
- Automated Trap Counting of Flea Beetles (FB), Aster Leafhoppers, and Wheat Midge and Automated FB Damage Assessment for CanolaSaskatchewan Ministry of Agriculture, Agriculture Development Fund2022–2025LeadPrincipal investigator
- Vehicle Re-identification using Computer Vision and Machine LearningMitacs, Accelerate2022PartnerCo-applicant
- Generative Face Models for Video Chat and BroadcastMitacs, Accelerate2022LeadPrincipal investigator
- Estimating Organic Matter, Soil Surface Roughness, Plant Stand Count, and Interplant Spacing from High-Resolution RGB ImageryMitacs, Accelerate2022LeadPrincipal investigator
- Laser Induced Breakdown Spectroscopy for In-Situ Soil AnalysisAlberta Innovates, Food Digitization and Automation Challenge2021–2023PartnerCo-applicant
- Ground Truth Validation of Crop Growth Cycle Using High Resolution Proximal and Remote SensingNSERC, Alliance Grant; Croptimistic Technology Inc., Alliance Grant2020–2023LeadPrincipal investigator
- Weed and Crop Leaf Area Estimation for Variable Rate ApplicationsMitacs, Accelerate2020–2021LeadPrincipal investigator
- Deployment of on-premises GPU based workstationUniversity of Regina, Engineering Research Opportunities Fund2020LeadPrincipal investigator
- Weed Classification and Density Estimation for Variable Rate Herbicide PrescriptionMitacs, Accelerate2019LeadPrincipal investigator
- Obstacle Detection and Avoidance in Remote Controlled Car Using Artificial IntelligenceMitacs, Globalink2019LeadPrincipal investigator
- Chronic Disease Prediction using Artificial IntelligenceMitacs, Globalink2019LeadPrincipal investigator
- Data Centric Protein Extraction Process OptimizationMitacs, Accelerate2018–2019LeadPrincipal investigator
- Academic Industry Meeting Day (AIMDay)-Ag Tech: Seed FundsInnovation Saskatchewan2018–2019LeadPrincipal investigator
- Academic Industry Meeting Day (AIMDay)-Big Data: Seed FundsInnovation Saskatchewan2018–2019LeadPrincipal investigator
- Deployment of On-premises Graphics Processing Units (GPU) Based Development Station for Deep LearningUniversity of Regina, Engineering Research Opportunities Fund2017LeadPrincipal investigator
- An Application Framework for Real-time Streaming Big Data AnalysisUniversity of Regina, SaskPower Clean Energy Research Grant2016–2019LeadPrincipal investigator
International collaborations
- King Fahd University of Petroleum & Minerals, Saudi ArabiaWith Prof. Aiman H. El-Maleh since 2024, on deep networks for camouflaged object detection.
- Edith Cowan University, AustraliaWith Dr. Syed Afaq Ali Shah since 2015, on deep learning for multimedia and social media analysis. We co-organized a workshop on deep learning for video and image analysis at PSIVT 2019 in Sydney.
- National University of Modern Languages, PakistanWith Dr. Sajid Saleem since 2010, on multisensor image matching, illumination-invariant feature descriptors and forensic speaker recognition.