LeafGuard: A Deep Learning System for Plant Leaf Disease Detection with Organic Treatment Recommendations

Farmers lose a huge chunk of their harvest to crop diseases every year, largely because getting a quick, dependable diagnosis out in the field is tough. To help with this, we built LeafGuard, a web application that identifies leaf infections from a photo and recommends organic ways to manage them. Our classifier is a MobileNetV2 network, initialized with ImageNet weights and fine-tuned on all 38 PlantVillage classes. Rather than outputting a bare class name, the site displays a Grad-CAM heat map so growers can see which spots the network focused on, and it uses a confidence cutoff to reject blurry or uncertain photos instead of forcing a bad guess. Every class links to a local JSON database of cultural habits, biological agents, and botanical or mineral remedies. On our test split, the model hit 98.4% accuracy and a 0.981 macro-F1 score, running in 42.6 ms per photo on a standard CPU. We also look closely at why clean lab photos from PlantVillage do not always translate directly to messy farm conditions.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-18
DOI
https://doi.org/10.5281/zenodo.24335726
Primary Topic
Smart Agriculture and AI
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

LeafGuard: A Deep Learning System for Plant Leaf Disease Detection with Organic Treatment Recommendations

Pradeep, Mohit Sharma
Zenodo (CERN European Organization for Nuclear Research)
Smart Agriculture and AI
article

LeafGuard: A Deep Learning System for Plant Leaf Disease Detection with Organic Treatment Recommendations

Pradeep, Mohit Sharma
article en

Abstract

Farmers lose a huge chunk of their harvest to crop diseases every year, largely because getting a quick, dependable diagnosis out in the field is tough. To help with this, we built LeafGuard, a web application that identifies leaf infections from a photo and recommends organic ways to manage them. Our classifier is a MobileNetV2 network, initialized with ImageNet weights and fine-tuned on all 38 PlantVillage classes. Rather than outputting a bare class name, the site displays a Grad-CAM heat map so growers can see which spots the network focused on, and it uses a confidence cutoff to reject blurry or uncertain photos instead of forcing a bad guess. Every class links to a local JSON database of cultural habits, biological agents, and botanical or mineral remedies. On our test split, the model hit 98.4% accuracy and a 0.981 macro-F1 score, running in 42.6 ms per photo on a standard CPU. We also look closely at why clean lab photos from PlantVillage do not always translate directly to messy farm conditions.

Zenodo (CERN European Organization for Nuclear Research)
Zero hunger
Smart Agriculture and AI
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon