First-author paper, IEEE ICCIT 2025 — DOI 10.1109/ICCIT68739.2025.11490422.

Problem & motivation

Whole-image plant-disease classifiers are easy to train and easy to fool: with cluttered field photos they happily learn background, soil, and lighting instead of the lesion. Farmers and agronomists also need a reason to trust a diagnosis, not just a label. This work addresses both with a two-stage design — find the leaf first, then classify the disease on what was found — and explainability overlays that show where the evidence is.

Approach

Stage 1 is a YOLO detector that localises the leaf region; stage 2 is an ECA-NFNet classifier operating on the detected crop; saliency overlays are generated for each prediction so the evidence can be visually checked against the symptoms. Cropping before classifying removes most background shortcut signals by construction.

Pipeline diagram: an input photo passes a YOLO leaf detector, the cropped leaf passes an ECA-NFNet disease classifier, and an explainability map is overlaid on the prediction
Detect, then classify, then show the evidence. (Placeholder — final figure from the paper.)

Results

Two-stage pipeline — headline results (IEEE ICCIT 2025)
StageMetricScore
Leaf detection (YOLO)mAP@0.5
Disease classification (ECA-NFNet)Top-1 accuracy (%)
TODO: fill in the published numbers from the paper (DOI 10.1109/ICCIT68739.2025.11490422).

Full quantitative results, ablations, and qualitative explainability examples are in the paper; code is on GitHub.

Demo

Try it on your own leaf photo

A Hugging Face Space (upload a photo → detection, diagnosis, and evidence overlay) is planned. TODO: add the Space URL.

Limitations & what I’d do next