| Title: | Superpixel Image Segmentation |
|---|---|
| Description: | Image Segmentation using Superpixels, Affinity Propagation and Kmeans Clustering. The R code is based primarily on the article "Image Segmentation using SLIC Superpixels and Affinity Propagation Clustering, Bao Zhou, International Journal of Science and Research (IJSR), 2013" <https://www.ijsr.net/archive/v4i4/SUB152869.pdf>. |
| Authors: | Lampros Mouselimis [aut, cre]
|
| Maintainer: | Lampros Mouselimis <[email protected]> |
| License: | GPL-3 |
| Version: | 1.0.5 |
| Built: | 2024-11-26 06:48:54 UTC |
| Source: | CRAN |
Segmentation of images based on Superpixels, Affinity propagation and Kmeans clustering
Segmentation of images based on Superpixels, Affinity propagation and Kmeans clustering
# init <- Image_Segmentation$new()# init <- Image_Segmentation$new()
sim_wL, sim_wA, sim_wB are the weights of the three channels. They keep balance so as to be consistent with human perception.
The quantity colorradius adjusts the number of clusters, and if its value is low, the number of targets would increase, which leads to more detailed segmentation results.
If the adjust_centroids_and_return_masks parameter is set to FALSE then the output kmeans_image_data will be an RGB image, otherwise it will be a black-and-white image.
colour_type parameter: RGB (Red-Green-Blue), LAB (Lightness, A-colour-dimension, B-colour-dimension) or HSV (Hue, Saturation, Value) colour.
Higher resolution images give better results.
The affinity propagation algorithm is used here with default parameter values.
By setting the sim_normalize parameter to TRUE, the affinity propagation algorithm requires less iterations to complete. However, the colorradius parameter does not have an effect if the similarity matrix is normalized.
Regarding the use_median parameter in the Rcpp I use the following steps: 1st. I compute the superpixels and extract the labels, 2nd. each superpixel label consists of multiple pixels and for these superpixels I have to compute a dissimilarity matrix therefore each superpixel must correspond to a single value, 3rd. to come to this single value for each superpixel the R user has the option to either use the 'mean' or the 'median of multiple image pixels (per superpixel)
—————kmeans initializers———————-
optimal_init : this initializer adds rows of the data incrementally, while checking that they do not already exist in the centroid-matrix
quantile_init : initialization of centroids by using the cummulative distance between observations and by removing potential duplicates
kmeans++ : kmeans++ initialization. Reference : http://theory.stanford.edu/~sergei/papers/kMeansPP-soda.pdf AND http://stackoverflow.com/questions/5466323/how-exactly-does-k-means-work
random : random selection of data rows as initial centroids
Image_Segmentation$new()--------------spixel_segmentation()--------------spixel_masks_show()--------------spixel_clusters_show()--------------new()
Image_Segmentation$new()
spixel_segmentation()
Image_Segmentation$spixel_segmentation( input_image, method = "slic", superpixel = 200, kmeans_method = "", AP_data = FALSE, use_median = TRUE, minib_kmeans_batch = 10, minib_kmeans_init_fraction = 0.5, kmeans_num_init = 3, kmeans_max_iters = 100, kmeans_initializer = "kmeans++", colour_type = "RGB", compactness_factor = 20, adjust_centroids_and_return_masks = FALSE, return_labels_2_dimensionsional = FALSE, sim_normalize = FALSE, sim_wL = 3, sim_wA = 10, sim_wB = 10, sim_color_radius = 20, ap_maxits = 1000, ap_convits = 100, ap_dampfact = 0.9, ap_nonoise = 0, verbose = FALSE )
input_imagea 3-dimensional input image (the range of the pixel values should be preferably in the range 0 to 255)
methoda character string specifying the superpixel method. It can be either "slic" or "slico"
superpixela numeric value specifying the number of superpixels
kmeans_methoda character string specifying the kmeans method. If not empty ("") then it can be either "kmeans" or "mini_batch_kmeans"
AP_dataa boolean. If TRUE then the affinity propagation image data will be computed and returned
use_mediana boolean. If TRUE then the median will be used rather than the mean value for the inner computations (see the details section for more information)
minib_kmeans_batchthe size of the mini batches
minib_kmeans_init_fractionpercentage of data to use for the initialization centroids (applies if initializer is kmeans++ or optimal_init). Should be a float number between 0.0 and 1.0.
kmeans_num_initnumber of times the algorithm will be run with different centroid seeds
kmeans_max_itersthe maximum number of clustering iterations
kmeans_initializerthe method of initialization. One of, optimal_init, quantile_init, kmeans++ and random. See details for more information
colour_typea character string specifying the colour type. It can be one of "RGB", "LAB" or "HSV"
compactness_factora numeric value specifying the compactness parameter in case that method is "slic"
adjust_centroids_and_return_masksa boolean. If TRUE and the kmeans_method parameter is NOT empty ("") then the centroids will be adjusted and image-masks will be returned. This will allow me to plot the masks using the spixel_masks_show method.
return_labels_2_dimensionsionala boolean. If TRUE then a matrix of labels based on the output superpixels in combination with the Affinity Propagation clusters will be returned
sim_normalizea boolean. If TRUE then the constructed similarity matrix will be normalised to have unit p-norm (see the armadillo documentation for more details)
sim_wLa numeric value specifying the weight for the "L" channel of the image (see the details section for more information)
sim_wAa numeric value specifying the weight for the "A" channel of the image (see the details section for more information)
sim_wBa numeric value specifying the weight for the "B" channel of the image (see the details section for more information)
sim_color_radiusa numeric value specifying the colorradius (see the details section for more information)
ap_maxitsa numeric value specifying the maximum number of iterations for the Affinity Propagation Clustering (defaults to 1000)
ap_convitsa numeric value. If the estimated exemplars stay fixed for convits iterations, the affinity propagation algorithm terminates early (defaults to 100)
ap_dampfacta float number specifying the update equation damping level in [0.5, 1). Higher values correspond to heavy damping, which may be needed if oscillations occur in the Affinity Propagation Clustering (defaults to 0.9)
ap_nonoisea float number. The affinity propagation algorithm adds a small amount of noise to data to prevent degenerate cases; this disables that.
verbosea boolean. If TRUE then information will be printed in the console (spixel_masks_show method)
spixel_masks_show()
Image_Segmentation$spixel_masks_show( delay_display_seconds = 3, display_all = FALSE, margin_btw_plots = 0.15, verbose = FALSE )
delay_display_secondsa numeric value specifying the seconds to delay the display of the next image (It displays the images consecutively). This parameter applies only if the display_all is set to FALSE (spixel_masks_show method)
display_alla boolean. If TRUE then all images will be displayed in a grid (spixel_masks_show method)
margin_btw_plotsa float number specifying the margins between the plots if the display_all parameter is set to TRUE (spixel_masks_show method)
verbosea boolean. If TRUE then information will be printed in the console (spixel_masks_show method)
spixel_clusters_show()
Image_Segmentation$spixel_clusters_show( spix_labels, color_palette = grDevices::rainbow, parameter_list_png = NULL )
spix_labelsa matrix. I can retrieve the "spix_labels" parameter by setting the "return_labels_2_dimensionsional" parameter to TRUE in the "spixel_segmentation" method (spixel_clusters_show method)
color_paletteone of the color palettes. Use ?grDevices::topo.colors to see the available color palettes
parameter_list_pngeither NULL or a list of parameters passed to the ?grDevices::png function, such as list(filename = 'img.png', width = 100, height = 100, units = "px", pointsize = 12, bg = "white", type = "quartz")
clone()
The objects of this class are cloneable with this method.
Image_Segmentation$clone(deep = FALSE)
deepWhether to make a deep clone.
https://www.ijsr.net/archive/v4i4/SUB152869.pdf , "Image Segmentation using SLIC Superpixels and Affinity Propagation Clustering", Bao Zhou, 2013, International Journal of Science and Research (IJSR)
library(SuperpixelImageSegmentation) path = system.file("images", "BSR_bsds500_image.jpg", package = "SuperpixelImageSegmentation") im = OpenImageR::readImage(path) init = Image_Segmentation$new() num_spix = 10 # for illustration purposes # num_spix = 600 # recommended number of superpixels spx = init$spixel_segmentation(input_image = im, superpixel = num_spix, AP_data = TRUE, use_median = TRUE, return_labels_2_dimensionsional = TRUE, sim_color_radius = 10) #........................... # plot the superpixel labels #........................... plt = init$spixel_clusters_show(spix_labels = spx$spix_labels, color_palette = grDevices::rainbow, parameter_list_png = NULL) # plt #.................................................... # create a binary image for a specified cluster label #.................................................... pix_values = spx$spix_labels target_cluster = 3 # determine clusters visually ('plt' variable) pix_values[pix_values != target_cluster] = 0 # set all other values to 0 (background) pix_values[pix_values == target_cluster] = 1 # set the target_cluster to 1 (binary image) # OpenImageR::imageShow(pix_values)library(SuperpixelImageSegmentation) path = system.file("images", "BSR_bsds500_image.jpg", package = "SuperpixelImageSegmentation") im = OpenImageR::readImage(path) init = Image_Segmentation$new() num_spix = 10 # for illustration purposes # num_spix = 600 # recommended number of superpixels spx = init$spixel_segmentation(input_image = im, superpixel = num_spix, AP_data = TRUE, use_median = TRUE, return_labels_2_dimensionsional = TRUE, sim_color_radius = 10) #........................... # plot the superpixel labels #........................... plt = init$spixel_clusters_show(spix_labels = spx$spix_labels, color_palette = grDevices::rainbow, parameter_list_png = NULL) # plt #.................................................... # create a binary image for a specified cluster label #.................................................... pix_values = spx$spix_labels target_cluster = 3 # determine clusters visually ('plt' variable) pix_values[pix_values != target_cluster] = 0 # set all other values to 0 (background) pix_values[pix_values == target_cluster] = 1 # set the target_cluster to 1 (binary image) # OpenImageR::imageShow(pix_values)