CocoDag — annotation tool
The tool that turns raw field imagery into labelled training data.
- bounding-box annotation
- labelled training data at scale
The challenge
Images from the coconut capture app needed bounding-box labels to train the yield-prediction pipeline — and there was no dedicated tagging tool.
What we built
CocoDag, an annotation tool wired directly to the capture app: images flow in, taggers draw bounding boxes around coconuts, labelled data flows out to the ML pipeline.
The outcome
Labelled training data produced at scale, feeding the yield-prediction models continuously.
Bring the problem. Leave with a scope, a timeline, and the number it should move.