kmeans.cpp File Reference

#include "kmeans.h"
#include "utils.h"

Functions
cv::Mat	kmeansBySize (const std::vector< cv::Mat > &extracted_templates, int K)
	Performs K-Means clustering on a set of templates based on their dimensions.
cv::Mat	kmeansByIntensity (const std::vector< cv::Mat > &clustered_templates_by_size, int K_clusters)
	Performs K-Means clustering on a set of templates based on their mean intensity.
void	saveClusteredImages (const std::vector< cv::Mat > &images, const std::vector< std::string > &image_paths, const cv::Mat &labels, const std::vector< std::filesystem::path > &cluster_paths)
	Saves images into directories based on their cluster labels.

Function Documentation

kmeansByIntensity()

cv::Mat kmeansByIntensity	(	const std::vector< cv::Mat > &	clustered_templates_by_size,
		int	K_clusters )

Performs K-Means clustering on a set of templates based on their mean intensity.

This function takes a vector of templates that have been clustered by size and performs K-Means clustering based on their mean intensity. The result is a matrix of labels indicating the cluster assignment for each template.

Parameters

[in]	clustered_templates_by_size	A vector of cv::Mat objects representing the templates that have been previously clustered by size.
[in]	K_clusters	The number of intensity-based clusters to form.

Returns

A cv::Mat containing the cluster labels for each template based on intensity.

Note

The function calculates the mean intensity of each template and uses this information as input for the K-Means clustering algorithm.

See also

cv::mean

cv::kmeans

kmeansBySize()

cv::Mat kmeansBySize	(	const std::vector< cv::Mat > &	extracted_templates,
		int	K )

Performs K-Means clustering on a set of templates based on their dimensions.

This function takes a vector of extracted templates and performs K-Means clustering based on their width and height. The result is a matrix of labels indicating the cluster assignment for each template.

Parameters

[in]	extracted_templates	A vector of cv::Mat objects representing the extracted templates.
[in]	K	The number of clusters to form.

Returns

A cv::Mat containing the cluster labels for each template.

Note

The function creates a matrix with the dimensions of each template and uses this matrix as input for the K-Means clustering algorithm.

See also

cv::kmeans

saveClusteredImages()

void saveClusteredImages	(	const std::vector< cv::Mat > &	images,
		const std::vector< std::string > &	image_paths,
		const cv::Mat &	labels,
		const std::vector< std::filesystem::path > &	cluster_paths )

Saves images into directories based on their cluster labels.

This function takes a vector of images, their corresponding file paths, cluster labels, and destination directories for each cluster. It saves each image into the appropriate directory based on its cluster label.

Parameters

[in]	images	A vector of cv::Mat objects representing the images to be saved.
[in]	image_paths	A vector of strings containing the original file paths of the images.
[in]	labels	A cv::Mat containing the cluster labels for each image.
[in]	cluster_paths	A vector of std::filesystem::path objects representing the destination directories for each cluster.

Note

The function assumes that the number of images, file paths, and labels are the same. Each image is saved with its original filename into the directory corresponding to its cluster label.

See also

cv::imwrite

std::filesystem::path