A structured pipeline, not one big guess
Splitting diagnosis into stages makes each one easier to train, test and explain. It also means the scanner can stop early when an image is not good enough.
Image check
Blur, exposure and framing
Sharp, well litCrop
Which of the twelve crops
MaizePlant part
Leaf, stem, pod or whole plant
LeafCondition
Ranked likely problems
Fall armyworm 92%Guidance
Curated next steps, spoken
In 14 languages
Trained on GPUs, run on the scanner
Models are trained centrally, then compressed to run fast on the handheld with no connection.
Collect field images
Images are captured with the scanner itself, so training data looks like what the model will see in use. Agronomists label each one.
Fine-tune on NVIDIA GPUs
Pretrained vision backbones are fine-tuned per crop, with balanced classes and augmentation for light, dust and leaf age.
Optimise for the edge
Models are quantised to 8-bit and compiled for the scanner’s neural processing unit, so a diagnosis takes under a second.
Validate by region
Each release is tested on held-out images from the region where it will be used, with results published per condition.
Ship as an update
Scanners download the release when connected and keep the previous version as a fallback.
Knowing when not to answer
Confidence is treated as part of the result. Low confidence changes what the scanner does next.
Image quality first
Blur, darkness and poor framing are caught before diagnosis, with a plain prompt to rescan.
Ranked alternatives
The top three options are always shown, with what to check in the field to tell them apart.
Escalation later
Unclear cases are queued and sent to an expert reviewer the next time the scanner connects.
Photo: Arnawakomang, CC BY-SA 4.0
Guidance is curated, never generated
The model decides what the problem probably is. What the farmer hears comes from a reviewed library written with local agronomists, so treatment advice stays appropriate and legal for each country.



