ViTaL: An Advanced Framework for Automated Plant Disease Identification in Leaf Images Using Vision Transformers and Linear Projection for Feature Reduction
2024 · Springer
A Vision Transformer pipeline for leaf disease ID, paired with affordable Raspberry Pi imaging hardware.
ViTaL is a plant-disease identification framework built around Vision Transformers, with linear-projection variants for feature reduction and CNN baselines for comparison. Pre-processing uses thumbnail resizing and normalization to keep compute light without discarding critical leaf detail.
On a PlantVillage subset of 3,306 images from six plant species, Architecture 2 without projection reports Hamming loss of 0.054. With linear projection it rises to 0.079. Lower is better: the comparison makes feature compression a measurable accuracy trade-off.
A proposed Raspberry Pi imaging rig provides the hardware context, alongside design registration 411009-001. The dataset experiments and proposed scanner should be read separately; the paper does not establish field performance of an integrated deployed device.
How does a model read a leaf?
Illustrated input
Patch values + where they belong
Extracted features → CNN classifier
Begin with the leaf.
A leaf image contains colour, texture, veins, and other visual patterns. This leaf is an original illustration, not a PlantVillage sample or a diagnosis.
Explore at your own pace.
Conceptual illustration inspired by Vision Transformer and the ViTaL paper. Patch count, colours, connections, and feature bars are simplified. No model runs here; no disease prediction is made.
- Vision transformers
- Agriculture
- Edge ML