Verdea AI

The project in context.

Verdea AI turns postharvest quality research into an image-analysis workflow that helps a user assess waterleaf freshness, compare deterioration across storage conditions and understand which physical indicators influenced the result.

A prediction still needs a useful product flow.

A freshness score alone does not explain what an image suggests or how quality may change under different storage conditions.

The product challenge was to translate model output into a result that remains approachable while retaining the physical indicators and comparisons behind it.

From image input to interpretable result.

• Image input: A focused upload flow captures the waterleaf image required for analysis.
• Analysis state: Loading and retry feedback keep the workflow understandable while the prediction request runs.
• Freshness result: A score and condition label give the user an immediate summary of the analysis.
• Comparative insight: Charts show projected freshness across ambient, refrigerated and frozen conditions.
• Explanation and guidance: Feature importance and storage recommendations add context beyond the headline result.

Make the prediction explainable in the interface.

The model result is presented as a structured product state rather than a single opaque answer. Freshness, deterioration comparisons and influential physical properties each explain a different part of the output.

This keeps machine learning in a supporting role: it produces the analysis, while the interface gives the user enough context to understand what the result means.

Keep the intelligence useful and bounded.

• Simple input vs useful context: Image upload keeps the interaction focused while the result carries the explanatory detail.
• Insight vs certainty: Scores and projections are presented with their contributing indicators rather than as unexplained facts.
• Responsiveness vs analysis time: Explicit loading and error states make the API boundary visible and recoverable.

What Verdea represents.

Verdea represents the product side of applied machine learning: turning research and model output into an experience that helps someone understand freshness, deterioration and storage choices more clearly.