Created
November 17, 2017 21:28
-
-
Save muschellij2/42525c0e8819a900133fce737671266d to your computer and use it in GitHub Desktop.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| ############################## | |
| # | |
| ############################## | |
| rm(list=ls()) | |
| library(matrixStats) | |
| library(RNifti) | |
| ensure_Nifti = function(x) { | |
| if (is.list(x)) { | |
| x = lapply(x, ensure_Nifti) | |
| } | |
| if (is.character(x)) { | |
| if (length(x) > 0) { | |
| x = lapply(x, readNifti) | |
| } else { | |
| x = readNifti(x) | |
| } | |
| } | |
| x | |
| } | |
| files = list.files( | |
| pattern = "Manual.*.nii.gz") | |
| # imgs = check_nifti(files) | |
| imgs = ensure_Nifti(files) | |
| orientations = sapply(imgs, orientation) | |
| ori = orientation(imgs[[1]]) | |
| imgs = lapply(imgs, | |
| function(x) { | |
| orientation(x) = ori | |
| x | |
| }) | |
| orig = t(sapply(imgs, c)) | |
| # bad = which(orig > 1, arr.ind = TRUE) | |
| orig = orig > 0 | |
| stopifnot(!any(is.na(orig))) | |
| keep = colSums(orig) | |
| stopifnot(!any(is.na(keep))) | |
| mat = orig[, keep > 0] | |
| umat = unique(c(mat)) | |
| f_t_i = colMeans(mat, na.rm = TRUE) | |
| dmat = 1 - mat | |
| mat[mat==0] = NA | |
| dmat[dmat==0] = NA | |
| n_readers = nrow(mat) | |
| n_voxels = ncol(mat) | |
| ################### | |
| #initialize | |
| p = q = rep(0.99999, n_readers) | |
| max_iter = 1000 | |
| tol = .Machine$double.eps | |
| # mat is D | |
| ### run E Step | |
| for (i in seq(max_iter)) { | |
| pmat = p * mat | |
| pmat = colProds(pmat, na.rm = TRUE) | |
| sep_pmat = (1-p) * dmat | |
| sep_pmat = colProds(sep_pmat, na.rm = TRUE) | |
| qmat = q * mat | |
| qmat = colProds(qmat, na.rm = TRUE) | |
| sep_qmat = (1-q) * dmat | |
| sep_qmat = colProds(sep_qmat, na.rm = TRUE) | |
| a_i = f_t_i * pmat * sep_pmat | |
| b_i = (1-f_t_i) * qmat * sep_qmat | |
| W_i = a_i/(a_i + b_i) | |
| sum_w = sum(W_i) | |
| new_p = t(mat) * W_i | |
| new_p = colSums(new_p, na.rm = TRUE) | |
| new_p = new_p/sum_w | |
| new_q = t(dmat) * (1 - W_i) | |
| new_q = colSums(new_q, na.rm = TRUE) | |
| new_q = new_q/(n_voxels - sum_w) | |
| diff_p = abs(p - new_p) | |
| diff_q = abs(q - new_q) | |
| diff = max(c(diff_p, diff_q)) | |
| if (diff <= tol) { | |
| print("Convergence!") | |
| break | |
| } else { | |
| print(paste0("iter: ", i, | |
| ", diff: ", diff)) | |
| } | |
| p = new_p | |
| q = new_q | |
| } | |
| stopifnot(!any(is.na(W_i))) | |
| outimg = rep(0, ncol(orig)) | |
| outimg[keep > 0] = W_i | |
| outimg = array(outimg, | |
| dim = dim(imgs[[1]])) | |
| hdr = RNifti::dumpNifti(imgs[[1]]) | |
| hdr$cal_max = 1 | |
| hdr$cal_min = 0 | |
| hdr$dattype = 16 | |
| outimg = RNifti::updateNifti( | |
| outimg, template = hdr) | |
| thresh = outimg >= 0.5 | |
| center = neurobase::xyz(thresh) | |
| flair = readNifti("3DFLAIR.nii.gz") | |
| orientation(flair) = ori | |
| neurobase::ortho2(flair, outimg, xyz = center) | |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| ############################## | |
| # | |
| ############################## | |
| rm(list=ls()) | |
| library(matrixStats) | |
| library(RNifti) | |
| ensure_Nifti = function(x) { | |
| if (is.list(x)) { | |
| x = lapply(x, ensure_Nifti) | |
| } | |
| if (is.character(x)) { | |
| if (length(x) > 0) { | |
| x = lapply(x, readNifti) | |
| } else { | |
| x = readNifti(x) | |
| } | |
| } | |
| x | |
| } | |
| files = list.files( | |
| pattern = "Manual.*.nii.gz") | |
| # imgs = check_nifti(files) | |
| imgs = ensure_Nifti(files) | |
| orientations = sapply(imgs, orientation) | |
| ori = orientation(imgs[[1]]) | |
| imgs = lapply(imgs, | |
| function(x) { | |
| orientation(x) = ori | |
| x | |
| }) | |
| orig = t(sapply(imgs, c)) | |
| # bad = which(orig > 1, arr.ind = TRUE) | |
| orig = orig > 0 | |
| stopifnot(!any(is.na(orig))) | |
| keep = colSums(orig) | |
| stopifnot(!any(is.na(keep))) | |
| mat = orig[, keep > 0] | |
| umat = unique(c(mat)) | |
| f_t_i = colMeans(mat, na.rm = TRUE) | |
| dmat = 1 - mat | |
| mat[mat==0] = NA | |
| dmat[dmat==0] = NA | |
| n_readers = nrow(mat) | |
| n_voxels = ncol(mat) | |
| ################### | |
| #initialize | |
| p = q = rep(0.99999, n_readers) | |
| max_iter = 1000 | |
| tol = .Machine$double.eps | |
| # mat is D | |
| ### run E Step | |
| for (i in seq(max_iter)) { | |
| pmat = p * mat | |
| pmat = colProds(pmat, na.rm = TRUE) | |
| sep_pmat = (1-p) * dmat | |
| sep_pmat = colProds(sep_pmat, na.rm = TRUE) | |
| qmat = q * mat | |
| qmat = colProds(qmat, na.rm = TRUE) | |
| sep_qmat = (1-q) * dmat | |
| sep_qmat = colProds(sep_qmat, na.rm = TRUE) | |
| a_i = f_t_i * pmat * sep_pmat | |
| b_i = (1-f_t_i) * qmat * sep_qmat | |
| W_i = a_i/(a_i + b_i) | |
| sum_w = sum(W_i) | |
| new_p = t(mat) * W_i | |
| new_p = colSums(new_p, na.rm = TRUE) | |
| new_p = new_p/sum_w | |
| new_q = t(dmat) * (1 - W_i) | |
| new_q = colSums(new_q, na.rm = TRUE) | |
| new_q = new_q/(n_voxels - sum_w) | |
| diff_p = abs(p - new_p) | |
| diff_q = abs(q - new_q) | |
| diff = max(c(diff_p, diff_q)) | |
| if (diff <= tol) { | |
| print("Convergence!") | |
| break | |
| } else { | |
| print(paste0("iter: ", i, | |
| ", diff: ", diff)) | |
| } | |
| p = new_p | |
| q = new_q | |
| } | |
| stopifnot(!any(is.na(W_i))) | |
| outimg = rep(0, ncol(orig)) | |
| outimg[keep > 0] = W_i | |
| outimg = array(outimg, | |
| dim = dim(imgs[[1]])) | |
| hdr = RNifti::dumpNifti(imgs[[1]]) | |
| hdr$cal_max = 1 | |
| hdr$cal_min = 0 | |
| hdr$dattype = 16 | |
| outimg = RNifti::updateNifti( | |
| outimg, template = hdr) | |
| thresh = outimg >= 0.5 | |
| center = neurobase::xyz(thresh) | |
| flair = readNifti("3DFLAIR.nii.gz") | |
| orientation(flair) = ori | |
| neurobase::ortho2(flair, outimg, xyz = center) | |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment