Meaning "foundation" or "where the source of waves begin" in Hawaiian.
This repository serves as the main codebase for training and inferencing the models that capture leaves and assess defect distribution.
More detailed information can be found at the website: Ainascan and at the following links:
- Instance Segmentation with Mask R-CNN on Coffee Leaves
- Tile-Based Semantic Segmentation of Coffee Leaves
The workflow.drawio file is the Draw.io diagram that shows the workflow of this project.
Open it to see how data is cataloged, annotated, trained, and inferenced.