reUse Grid stitching generated coords
Locked
reUse Grid stitching generated coords
 
|
Re: reUse Grid stitching generated coords
|
|
Re: reUse Grid stitching generated coords
|
|
Re: reUse Grid stitching generated coords
|
|
Re: reUse Grid stitching generated coords
|
|
In reply to this post by simone codeluppi
You can apply the TileConfiguration transformation to other data, as
long as the TEXT is made to match the new data. We do this on big data that would not all fit into RAM, so the case below is for input data from individual tifs with the common XY_Z_C_etc naming scheme. This macro code works on Leica exported data and illustrates how to go about it. [WARNING pasted code is likely to have long lines split; best to use attached IJM (if it makes it as attachment)]. The Z matching option seems to give incorrect results for our data. The default linear blend fusion is likewise often too obvious, so I favor the 'random intensity' fusions. The tolerances for accepting tiles are deliberately set low so as to prevent loss of tiles with little data and in a grid of known layout. You can use the macro recorder to generate other fusion options (an extremely long text line!), and insert as needed. __________macro start cut__________________________________________ /*Macro to apply a common transform to align and stitch 3D mosaics using FIJI Grid/Collection plugin * Specific case is for Leica LIF data, using a specific channel as a reference; for BIG DATA! * Since bioformats fails to correctly open LIF tile scanning order, this macro requires export to single tifs * and pulls in all stage/CH/Z files for stitching. PREP WORK: * Open a Leica merge to count up the RxC values used in capture * EXPORT_AS_TIF the LIF file into a folder using Leica software, and create a separate output folder * Look at last export files to get Z step numbers, BUT, ADD ONE [starts with zero not 1] * RefCH is main reference for all others --fit performed for RefCH and then same is applied to all other channels * and then assembles all into a composite stack [if memory allows] * This macro EDITS the transformation text file to call each channel (so you don't have to manually!) serially * You can edit the code below for other stitching options; default is to ignore Z shifts and use random intensity blending [best options!] * <v8 bug fixes: input dimensions, plus max projection generation at end * Code is for current FIJI, all features as default plugins/features * this pads names for grid stitcher with leading zeros to 999 stage locales * V.Bindokas, Univ of Chicago, JULY 2015 */ run("Close All"); Dialog.create("setup"); Dialog.addNumber("tiles across", 5); Dialog.addNumber("vertical", 7); Dialog.addNumber("Ch", 4); Dialog.addNumber("Z", 3); Dialog.addNumber("RefCH", 1); Dialog.show(); Xs = Dialog.getNumber()"Rows"; //order read is order above! Ys = Dialog.getNumber()"Col"; ch = Dialog.getNumber()"Ch"; z = Dialog.getNumber()"Z"; st = Xs * Ys; ref= Dialog.getNumber()"RefCH"; d1=getDirectory("Choose input Directory"); li=getFileList(d1); d2=getDirectory("Choose output Directory"); fup=d1+li[1]; for (S=0; S<st; S++){ //this makes blocks of images the tiler seems to require if (st>9 && S<10) ss="0"+toString(S); //stringify and pad counter value to match leica naming scheme [number of digits] else ss=toString(S); if (st>99 && S<100 ) ss="0"+ss; //stringify and pad, add another zero for the big-job folks... run("Image Sequence...", "open=[fup] file=_s"+ss+" sort"); //import all files per stage position run("Stack to Hyperstack...", "order=xyczt(default) channels=&ch slices=&z frames=1 display=Color"); run("Gaussian Blur...", "sigma=1 stack"); //denoise all, just a touch run("Split Channels"); for (h=ch;h>=1;h--){ pa=d2+"S"+ss+"_C"+h+".tif"; saveAs("Tiff", pa); //write out data for stitching, keeping memory needs lower close(); } //for h } //for S //recorded stitch setup below basically keeps every tile sans 'calculation' of fits to avoid the common goof ups dropping things or shifting Z for no good end //change to whatever you feel like trying. The alignment map is then applied to all other channels vs trying to fit each independently. So, make sure reference CH is a good one! if (st>9 && st<99){ run("Grid/Collection stitching", "type=[Grid: row-by-row] order=[Right & Up] grid_size_x=&Xs grid_size_y=&Ys tile_overlap=10 first_file_index_i=0 directory=[&d2] file_names=S{ii}_C"+ref+".tif output_textfile_name=TileConfiguration.txt fusion_method=[Intensity of random input tile] regression_threshold=0.00 max/avg_displacement_threshold=115 absolute_displacement_threshold=333 ignore_z_stage subpixel_accuracy display_fusion computation_parameters=[Save computation time (but use more RAM)] image_output=[Fuse and display]"); } if (st>99 ){ run("Grid/Collection stitching", "type=[Grid: row-by-row] order=[Right & Up] grid_size_x=&Xs grid_size_y=&Ys tile_overlap=10 first_file_index_i=0 directory=[&d2] file_names=S{iii}_C"+ref+".tif output_textfile_name=TileConfiguration.txt fusion_method=[Intensity of random input tile] regression_threshold=0.00 max/avg_displacement_threshold=115 absolute_displacement_threshold=333 ignore_z_stage subpixel_accuracy display_fusion computation_parameters=[Save computation time (but use more RAM)] image_output=[Fuse and display]"); } else { run("Grid/Collection stitching", "type=[Grid: row-by-row] order=[Right & Up] grid_size_x=&Xs grid_size_y=&Ys tile_overlap=10 first_file_index_i=0 directory=[&d2] file_names=S{i}_C"+ref+".tif output_textfile_name=TileConfiguration.txt fusion_method=[Intensity of random input tile] regression_threshold=0.00 max/avg_displacement_threshold=115 absolute_displacement_threshold=333 ignore_z_stage subpixel_accuracy display_fusion computation_parameters=[Save computation time (but use more RAM)] image_output=[Fuse and display]"); } pa="fused_ch"+ref+".tif"; saveAs("Tiff", d2+pa); close(); dx = newArray("1"); //build array for doing other channels if (ch>1){ for (dxi=2; dxi<=ch; dxi++){ dx=Array.concat(dx, dxi); //Array.print(dx);// } for (i=0;i<ch; i++){ // BE AWARE array index starts as ZERO not 1 if ( parseInt(dx[i]) != ref){ A="_C"+toString(ref); B="_C"+toString(parseInt(dx[i])); nu=d2+"next.txt"; //register data map with next CH in names vs CH<ref> of 1st run fi=d2+"TileConfiguration.txt"; str=File.openAsString(fi); rows=split(str, "\n"); x=newArray(rows.length); for(k=0; k<rows.length; k++){ //do the filename string edits for new input rows[k]=replace(rows[k], A, B); print(rows[k]); File.append(rows[k], nu); File.append("\r", nu); } //apply stitiching and write result to disk run("Grid/Collection stitching", "type=[Positions from file] order=[Defined by TileConfiguration] directory=[&d2] layout_file=next.txt fusion_method=[Intensity of random input tile] regression_threshold=0.00 max/avg_displacement_threshold=115 absolute_displacement_threshold=333 ignore_z_stage subpixel_accuracy display_fusion computation_parameters=[Save computation time (but use more RAM)] image_output=[Fuse and display]"); pa="fused_ch"+(dx[i])+".tif"; saveAs("Tiff", d2+pa); File.delete(nu); close(); } //end if not ref CH } //end for all CH } //end ch>1 //pull all CH output back in for building the large composite for (i=1;i<=ch; i++){ run("Bio-Formats Importer", "open=["+d2+"fused_ch"+i+".tif] color_mode=Default view=Hyperstack stack_order=XYCZT"); //virtual stack has bug--- not doing Z! } Stack.getDimensions(width, height, channels, slices, frames); run("Concatenate...", "all_open title=all"); run("Stack to Hyperstack...", "order=xyzct channels=&ch slices=&slices frames=1 display=Composite"); //you can add steps to set colors specifically after this line, if you need to pa="fused_merged.tif"; saveAs("Tiff", d2+pa); run("Z Project...", "projection=[Max Intensity]"); __________macro end____________________ On 8/13/2015 9:57 AM, simone codeluppi wrote: > Hi > My goal is to stitch a multi-channels image using the coordinates or > overlapping of the tiles generated in the stitching of a reference channel > (ex. DAPI staining). I first run the plugin stitching the Reference channel > (grid setting) saving the TileConfiguration or TileConfiguration.registered > file. I then I re-run the plugin on a different channel using the "Position > from file" setting from TileConfiguration or TileConfiguration.registered > but the stitching result is slightly different. Is there a way to "force" > the plugin to place the tiles using the previously generated coord? > > Any help will be greatly appreciated. > Thanks a lot! > > Simone > > > > > > -- > View this message in context: http://imagej.1557.x6.nabble.com/reUse-Grid-stitching-generated-coords-tp5014013.html > Sent from the ImageJ mailing list archive at Nabble.com. > > -- > ImageJ mailing list: http://imagej.nih.gov/ij/list.html __ Vytas Bindokas, Ph.D. Research Assoc. (Assoc. Prof.), Director, BSD Light Microscopy Core Facility phone: 773-702-4875 [address for letters ONLY (see shipping addr below):] Dept Pharmacol Physiol Sci MC0926 947 E 58th Street The University of Chicago Chicago IL 60637 Room Abbott 129 shipping address (main KCBD site): V. Bindokas 900 E 57th Street KCBD room 1250, Microscopy Core The University of Chicago Chicago IL 60637 email [hidden email] web site for LMCF: http://digital.uchicago.edu/index.html -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
LIFtile_restitcher v8.ijm (6K)
Re: reUse Grid stitching generated coords
|
|
Re: reUse Grid stitching generated coords
|
|
| Free forum by Nabble | Edit this page |