glcm: Creates gray-level co-occurrence matrix of RIA image

In neuroconductor/RIA: Radiomics Image Analysis Toolbox for Medial Images

Description

Creates gray-level co-occurrence matrix (GLCM) from RIA_image. GLCM assesses the spatial relation of voxels to each other. By default the $modif image will be used to calculate GLCMs. If use_slot is given, then the data present in RIA_image$use_slot will be used for calculations. Results will be saved into the glcm slot. The name of the subslot is determined by the supplied string in save_name, or is automatically generated by RIA.

Usage

glcm(RIA_data_in, off_right = 1, off_down = 0, off_z = 0,
  symmetric = TRUE, normalize = TRUE, use_type = "single",
  use_orig = FALSE, use_slot = NULL, save_name = NULL,
  verbose_in = TRUE)

Arguments

RIA_data_in	RIA_image.
off_right	integer, indicating the number of voxels to look to the right. Negative values indicate to the left.
off_down	integer, indicating the number of voxels to look down. Negative values indicate up.
off_z	integer, indicating the number of voxels to look in cross plane.
symmetric	logical, indicating whether to create a symmetric glcm by also calculating the glcm in the opposite direction (-1*off_right; -1*off_down; -1*off_z), and add it to the glcm
normalize	logical, indicating whether to change glcm elements to relaive frequencies.
use_type	string, can be "single" which runs the function on a single image, which is determined using "use_orig" or "use_slot". "discretized" takes all datasets in the RIA_image$discretized slot and runs the analysis on them.
use_orig	logical, indicating to use image present in RIA_data$orig. If FALSE, the modified image will be used stored in RIA_data$modif.
use_slot	string, name of slot where data wished to be used is. Use if the desired image is not in the data$orig or data$modif slot of the RIA_image. For example, if the desired dataset is in RIA_image$discretized$ep_4, then use_slot should be discretized$ep_4. The results are automatically saved. If the results are not saved to the desired slot, then please use save_name parameter.
save_name	string, indicating the name of subslot of $glcm to save results to. If left empty, then it will be automatically determined by RIA.
verbose_in	logical indicating whether to print detailed information. Most prints can also be suppressed using the suppressMessages function.

Value

RIA_image containing the GLCM.

References

Robert M. HARALICK et al. Textural Features for Image Classification. IEEE Transactions on Systems, Man, and Cybernetics. 1973; SMC-3:610-621. DOI: 10.1109/TSMC.1973.4309314 http://ieeexplore.ieee.org/document/4309314/

Márton KOLOSSVÁRY et al. Radiomic Features Are Superior to Conventional Quantitative Computed Tomographic Metrics to Identify Coronary Plaques With Napkin-Ring Sign Circulation: Cardiovascular Imaging (2017). DOI: 10.1161/circimaging.117.006843 https://www.ncbi.nlm.nih.gov/pubmed/29233836

Márton KOLOSSVÁRY et al. Cardiac Computed Tomography Radiomics: A Comprehensive Review on Radiomic Techniques. Journal of Thoracic Imaging (2018). DOI: 10.1097/RTI.0000000000000268 https://www.ncbi.nlm.nih.gov/pubmed/28346329

Examples

## Not run: 
#Discretize loaded image and then calculate GLCM matrix of RIA_image$modif
RIA_image <- discretize(RIA_image, bins_in = c(4, 8), equal_prob = TRUE,
use_orig = TRUE, write_orig = FALSE)
RIA_image <- glcm(RIA_image, use_orig = FALSE, verbose_in = TRUE)

#Use use_slot parameter to set which image to use
RIA_image <- glcm(RIA_image, use_orig = FALSE, use_slot = "discretized$ep_4",
off_right = 2, off_down = -1, off_z = 0)

#Batch calculation of GLCM matrices on all discretized images
RIA_image <- glcm(RIA_image, use_type = "discretized",
off_right = 1, off_down = -1, off_z = 0)

## End(Not run)

neuroconductor/RIA documentation built on May 21, 2021, 6:56 a.m.