Configuration

Configuration

Project structure

This project is composed by 3 Python packages, each one containing a script for: k-means for initial segmentation, merging algorithm, and image classification. The program executes these scripts sequentially from main.py. The file config.yaml contains the configuration regarding directories used during the execution and storing of results.

config.yaml structure:

The file is self-explanatory; it is divided into the following key-values structure. The directories are default configured but you could modify these settings to your needs.

key	Type	Description
dev_mode	Bool	Operation mode; false to be used with unlabeled segments and true with labeled segments. The later is used when you want to develop: measure performance at each step, create new features or train a new model using a supervised classifier.
save_dataset	Bool	Use this option to keep copies of your datasets between runs. Watch your storage capacity between runs.
k_means_segmentation	String	All the parameters related to k-means tuning and directories to store results.
merge_algorithm	String	All the parameters related to the merging algorithm tuning and directories to store results
model	String	The paths to the model location and region output.