This is the official repository to the paper "Problem-oriented Indicators for Semantic Segmentation and Determining Number of Plants" by Paweł Majewski and Jacek Reiner.
Datasets are available under the link https://drive.google.com/drive/folders/13ZJbOUWGAEkRtd9LJnYEJrmyercfWyOq?usp=sharing.
Put data folder in the same location as *.py files
- add_color_spaces_testing.py - Training and testing models with additional channels from HSV and Lab color spaces
- counting_plants.py - Training and testing models for plants counting
- create_chart_f1_score_vs_no_samples.py - Creating charts f1-score vs number of labeled samples
- error_VARI_bar_plot.py - Creating bar chart to show estimation errors of mean VARI for plants
- f1_score_vs_no_samples.py - Training and testing models for segmentation with a different number of labeled samples.
- fraction_indicator_PSO_optimatization.py - PSO implementation for fraction indicators optimatization
- mean_VARI_segmentation.py - Calculating estimation errors of mean VARI for plants
- opt_parameters.py - Optimized parameters for chosen models
- ROC_segmentation.py - Creating ROC-curves as threshold-indepentednt analysis for segmentation
- segmentation_plants.py - Training and testing models for plants segmentation
- universality.py - Checking universality of chosen models
- utilities.py - Universal functions used in implementation
- visualization_segmentation.py - Visualization of segmentation for chosen models