mvallieres / radiomics Goto Github PK
View Code? Open in Web Editor NEWMATLAB programming tools for radiomics analysis
License: GNU General Public License v3.0
MATLAB programming tools for radiomics analysis
License: GNU General Public License v3.0
___ ____ _____ ____ _ _ _____ ___ __ | | /\ | \ | / \ | \ / | | / \ / | |___| / \ | | | | | | \/ | | | \__ | \ /____\ | | | | | | | | | \ | \ / \ |____/ __|__ \____/ | | __|__ \___/ |__/ ------------------------------------------------------------------------- |<https://github.com/mvallieres/radiomics/>| --> A package providing MATLAB programming tools for radiomics analysis. ------------------------------------------------------------------------- REFERENCES: [1] Vallières, M. et al. (2015). A radiomics model from joint FDG-PET and MRI texture features for the prediction of lung metastases in soft-tissue sarcomas of the extremities. Physics in Medicine and Biology, 60(14), 5471-5496. doi:10.1088/0031-9155/60/14/5471 [2] Zhou, H., Vallières, M., Bai, H.X. et al. (2017). MRI features predict survival and molecular markers in diffuse lower-grade gliomas. Neuro-Oncology, 19(6), 862-870. doi:10.1093/neuonc/now256 [3] Vallière, M. et al. (2017). Radiomics strategies for risk assessment of tumour failure in head-and-neck cancer. Scientific Reports, 7:10117. doi:10.1038/s41598-017-10371-5 ------------------------------------------------------------------------- AUTHOR: Martin Vallières <[email protected]> ------------------------------------------------------------------------- HISTORY: - Version 1.0: May 2015 ------------------------------------------------------------------------- DISCLAIMER: "I'm not a programmer, I'm just a scientist doing stuff!" ------------------------------------------------------------------------- *** THANK YOU FOR YOUR INTEREST IN THIS PACKAGE *** --> If you have any questions, comments or suggestions about this package, please do not hesitate to contact me! This package contains 5 folders: 1. 'TextureToolbox': MATLAB codes to perform texture analysis from an input 2D or 3D region of interest (ROI). This toolbox is self-contained and can be used on its own outside of the radiomics package. In particular, this texture analysis package implements wavelet band-pass filtering, isotropic resampling, discretization length corrections and different quantization tools. Please see ref. [1] for more details. 2. 'NonTextureFeatures': MATLAB codes to compute features other than textures from an input 3D region of interest (ROI). Include features such as SUV metrics, AUC-CSH, Percent Inactive, Size, Solidity, Volume and Eccentricity. Please see ref. [1] for more details. 3. 'MultivariableModeling': MATLAB codes to perform multivariable analysis operations such as logistic regression, bootstrapping, feature set reduction, feature set selection, prediction performance estimation, etc. 4. 'Utilities': MATLAB codes used to perform different operations including the computation of SUV maps, reading of directory containing DICOM imaging data, conversion of RTstruct DICOM objects to 3D masks, etc. 5. 'STUDIES': MATLAB codes used for specific studies. To reproduce the experiments of a given study, please see its corresponding folder. - 'STS_study': Ref. [1] A. Imaging/ROI data and clinical information is available on The Cancer Imaging Archive (TCIA) website under the following DOI: <http://dx.doi.org/10.7937/K9/TCIA.2015.7GO2GSKS>. - 'LGG_study': Ref. [2] A. Texture data ("TEXTURES_TCGA.zip") is available on Google Drive: <https://drive.google.com/open?id=0B0fcZCGXT3nZWXM5d0t3OXVjQzA> B. Imaging data is available on the TCIA website: <http://doi.org/10.7937/K9/TCIA.2016.L4LTD3TK> C. ROI data is available on the TCIA website: <https://doi.org/10.7937/K9/TCIA.2017.BD7SGWCA> - 'HN_study': Ref. [3] A. Imaging/ROI data and clinical information is available on the TCIA website: <http://doi.org/10.7937/K9/TCIA.2017.8oje5q00> *************************************************************************** ACKNOWLEDGEMENTS: other software code - Wei's GLRLM toolbox: Xunkai Wei, Gray Level Run Length Matrix Toolbox v1.0, Software,Beijing Aeronautical Technology Research Center, 2007. <http://www.mathworks.com/matlabcentral/fileexchange/17482-gray-level-run-length-matrix-toolbox> - Q. Li: <http://www.mathworks.com/matlabcentral/fileexchange/23377-ellipsoid-fitting> - CERR development team: <http://www.cerr.info/> - Dirk-Jan Kroon (imresize3D.m): <http://www.mathworks.com/matlabcentral/fileexchange/21451-multimodality-non-rigid-demon-algorithm-image-registration/content//functions/imresize3d.m> - David Reshef and Yakir Reshef: MINE version 1.0.1d <http://www.exploredata.net/> - DREES development team: <http://www.cerr.info/drees> - Enric Junqué de Fortuny (fastAUC.cpp): <http://www.mathworks.com/matlabcentral/fileexchange/41258-faster-roc-auc> - François Beauducel (roundsd.m): <http://www.mathworks.com/matlabcentral/fileexchange/26212-round-with-significant-digits> - Jos van der Geest (herrorbar.m): <http://www.mathworks.com/matlabcentral/fileexchange/3963-herrorbar> ***************************************************************************
Is it run directly? I can not run it with matlab 2017b,which just shows a figure without functions. So,any idea?
Hi, thanks for posting the data out for everyone to access. I downloaded the .mat files containing the radiomics features. I noticed that the four modalities combined total upto 408 files. Out of these 106 are in T1CE, 107 are in T1W, 108 in T2W, and 87 in T2F. Is there any reason behind the difference in the number of mat files in the four modalities. This also doesn't tally up with the info mentioned [here]. (https://github.com/mvallieres/radiomics/blob/master/STUDIES/LGG_study/TCGA_clinical_INFO.csv)
Hi. first of all, thank you so much .
i tried to test the code with samples.
I got 44 features.
I'd like to add wavelet based features : e.g. multi-scale textures of variance - five features: LHL, HHL, LLH, LHH, and HHH), entropy (i.e., five features: HLL, LHL, HHL, LLH and HHH), and energy (i.e., two features: HLL and LLH). ...
What i want is all features : 44 * (5+5+2) features.
How can I get all features ?
thank you.
Based on this question (https://stackoverflow.com/questions/59308919/radiomics-package-in-matlab)
I used is this code:
volume = double(rgb2gray(imread('http://www.cs.sjsu.edu/~bruce/images/fall_2016_cs160/lectures/eye_pupil_tracking/grayscale_eye_cropped_to_bounding_box.png')));
mask = ones(size(volume,1),size(volume,2));
[ROIonly,levels] = prepareVolume(volume,mask,'Other',10, 22, 1, 22,'Matrix','Uniform',256);
[GLCM] = getGLCM(ROIonly,levels);
DataGLCM = getGLCMtextures(GLCM)
I do not know what is the pixelW so what will be the value? In the case of this png file what shell I use as "Numerical value specifying the in-plane resolution (mm) of 'volume'"? (I used a random number of 10), What shell I used in scale as it is related to pixelW?. When I will use it on eye images I will use the full image like in the case above where mask and volume are the same.
Thanks a lot,
Lluis
Hi mvallieres!
I have to process some fundus images and i don't know the number of pixelW and sliceS..... So how to solve this problem?? Thanks!
Hi mvallieres!
Thanks for your codes.
When I use the function "prepareVolume" with a 2D data, I got a error like:
error wavedec3 (line 109)
sizes(level+1,1:3) = size(X);
error waveletBPfilt (line 51)
wDEC=wavedec3(volume,1,wavelet);
error prepareVolume (line 149)
ROIbox = waveletBPfilt(ROIbox,R,'sym8');
I think that's because the 'size(X)' has 2 , while the 'sizes(level+1,1:3)' has 3. You said the function "prepareVolume" is compatible with 2D analysis. But it looks doesn't work for 2D analysis.
How could I revise it? Thanks!
I run LGG_study masterScript_LGG on OS matlab_R17a run ,as your INSTRUCTIONS,it goes on well ,untill
% A.2 PERFORM FEATURE SET REDUCTION FROM TRAINING SETS
tic, fprintf('\n--> PERFORMING FEATURE SET REDUCTION FROM THE TRAINING SETS WITH %u CORES ... ',nBatch)
calcAllFeatureSets_batchLGG(pathMINE,pathLR,setSize,alpha,delta,nBoot,nBatch,matlabPATH,seeds(1))
Matlab Window:
--> PERFORMING FEATURE SET REDUCTION FROM THE TRAINING SETS WITH 8 CORES ... Undefined function or variable 'batchExperiments'.
Error in calcAllFeatureSets_batchLGG (line 82)
[param] = batchExperiments(setNames,outcomes,nBatch); nBatch = length(param);
Could you please give me a help ?
thanks!!!
Ubuntu 16.04
R2015b
--> PERFORMING PREDICTION PERFORMANCE ESTIMATION FROM THE TRAINING SETS WITH 8 CORES ... DONE!
时间已过 66.557156 秒。
错误使用 load
无法读取文件 'RESULTS_T1W_T2W_nonIDH1'。没有此类文件或目录。
出错 plotPredictionResults_LGG (line 70)
results = load(['RESULTS_',fSET,'_',nameOutcome]); results = struct2cell(results); results =
results{1};
出错 autoChoiceBestModels_LGG (line 66)
plotPredictionResults_LGG(fullfile(pathExperimentBinary,'RESULTS'),nameOutcomes{o},fSetNames,metric,maxOrder,pathFig)
出错 masterScript_LGG (line 190)
autoChoiceBestModels_LGG(pathLR,fSetNames,nameOutcomes,'AUC632',maxOrder,pathFig) % USE THIS LINE
TO REPRODUCE THE RESULTS OF THIS STUDY.
Hi @mvallieres I try to run the masterScript_LGG.m. I have followed the instruction in README in LGG_study folder step by step.But error just happened as above shown.Do you know what is the problem?Thanks.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.