sample_config.yml

# unsigned integer
slic_size: 35

# float between 0 ad 1
slic_regul: 0.2

# OPTIONS: {
#	color: ['mean', 'std', 'energy', 'median', 'meanGrad'],
#	color_hsv: ['mean', 'std', 'energy', 'median', 'meanGrad'],
#	color_luv: ['mean', 'std', 'energy', 'median', 'meanGrad'],
#	color_lab: ['mean', 'std', 'energy', 'median', 'meanGrad'],
#	color_hed: ['mean', 'std', 'energy', 'median', 'meanGrad'],
#	color_xyz: ['mean', 'std', 'energy', 'median', 'meanGrad'],
#	tLM:   ['mean', 'std', 'energy', 'median', 'meanGrad'],
#	tLM_short: ['mean', 'std', 'energy', 'median', 'meanGrad']
# }
features:
  color_rgb: ['mean', 'std', 'energy']

# float between 0 and 1 or None
pca_coef: null

# OPTIONS: ['GMM', 'GMM_kmeans', 'GMM_Otsu', 'kmeans', 'kmeans_quantiles', 'BGM']
estim_model: 'GMM'

# float between 0 ad 1
label_purity: 0.95

# OPTIONS: ['None', 'random', 'kmeans', 'unique']
balance: 'unique'

# OPTIONS: ['RandForest', 'KNN', 'SVM', 'DecTree', 'GradBoost', 'LogistRegr', 'AdaBoost']
classif: 'RandForest'

# OPTIONS: ['f1', 'accuracy', 'precision', 'recall']
classif_metric: 'f1'

# unsigned integer
nb_classif_search: 50

# OPTIONS: ['spatial', 'color', 'features', 'model', 'model_l1', 'model_l2', 'model_lT']
gc_edge_type: 'model'

# list of labels to be skipped
drop_labels: []

# positive float
gc_regul: 3.0

# float between 0 and 1
cross_val: 0.1

run_LPO: true
run_groupGMM: true