iota2.validation.results_utils

utility functions for report generation

Functions

compute_interest_matrix(all_matrix[, f_interest])	from a list of confusion matrix compute the mean of each confusion
concatenate_sample_selection(result_folder, ...)	concatenates the csv files produces by sampleselection step result_folder: folder where output csv will be written nomenclature_path: file where labels nomenclature is defined outrates_path: path of the OTB generated outrates.csv files seed: number of the seed regions: set of region numbers labels_conversion: maps i2-encoded label to original label
fig_conf_mat(conf_mat_dic, nom_dict, kappa, ...)	usage : generate a figure representing the confusion matrix
gen_confusion_matrix_fig(csv_in, out_png, ...)	usage : from a csv file representing a confusion matrix generate the corresponding figure
get_coeff(matrix)	use to extract coefficients (Precision, Recall, F-Score, OA, K) from a confusion matrix.
get_color_table(color_file)	usage parse color file
get_conf_max(conf_mat_dic, nom_dict)	get confusion max by class
get_interest_coeff(runs_coeff, nb_lab[, ...])	use to compute mean coefficient and 95% confidence interval.
get_max_labels(conf_mat_dic, nom_dict)	return the maximum len of all labels
get_nomenclature(nom_path)	usage parse nomenclature file
get_region_and_seeds(path)	search path to find the regions and seeds of outrates files
get_rgb_mat(norm_conf, diag_cmap, not_diag_cmap)	usage convert normalise confusion matrix to a RGB image
get_rgb_pre(pre_val, coeff_cmap)	convert values to a RGB code thanks to a colormap
get_rgb_rec(coeff, coeff_cmap)	usage convert normalise confusion matrix to a RGB image
merge_all_sample_selection(output_path, ...)	searches for all existing sample selections csv files and for each seed, merges them in a single file output_path: output path root containing learningSamples folder and sampleSelections folders nomenclature_path: file where labels nomenclature is defined labels_conversion: maps i2-encoded label to original label
merge_class_matrix(conf_mat_dic, merge_class)	param conf_mat_dic a confusion matrix (index represent row reference)
normalize_conf(conf_mat_array[, norm])	function use to normalize a numpy array representing a confusion matrix with ref as row and production in column
parse_csv(csv_in)	parse OTB's confusion matrix
remove_undecidedlabel(conf_mat_dic, ...)	usage : use to remove samples with the undecidedlabel label from the confusion matrix
stats_report(csv_in, nomenclature_path, ...)	usage : sum-up statistics in a txt file

Classes

OrderedDict

Dictionary that remembers insertion order