Skip to content
Snippets Groups Projects
Select Git revision
  • master default
You can move around the graph by using the arrow keys.
Created with Raphaël 2.2.028Mar241715987622Feb2118Jan930Dec1917Nov1019Aug23Jun26Apr1918151413765431Fix issue in measure_dense_phases where dense phase masks were the wrong shape and thus giving the wrong mean/SD values for the dense phase intensities. Tested on one image and values look correct.mastermasterFix issue with color channel axes in processMFLLPS.py.Re-write analysis code downstream of creating the droplet label image. Tested on a single image and works.Add progress bar to makeimagestack.py.Fix error in makehistogram.py for single word color channel names. Change makeimagestack.py to handle areas (a.k.a. ROIs) correctly. Changes in rfsegment.py and rftrain.py after testing.Add option for logscale y-axis to makehistogram.py.Revert changes in get_leica_xml_info() from previous commit and handle tilescan stage positions within main(). Tested and tilescan stage coordinates look correct.Start modifying lif2tif.py to add stage position metadata to appropriate tag for use with multiscale_segmentation.py.Initial commit for multiscale_segmentation.py.Implement random forest model for multi-exposure image stacks (rftrain_imgx.py) and other minor changesImplement saving into specified output directory.Fix default value for command line args.Initial commit. Full functionality but not tested versus ImageJ/FIJI yet.Fix bug that was causing Z-slices to be out of order.Implement arbitrary numbers of color channels and Z positions.Fix error in argument names.Add option for mean global thresholding.Make dark subtraction and flat field correction optional.Implement dark image subtraction and flat field correction.Add handling 2D images in LoG blob detection calculation.Change output condensate_masks_color.tif so that three color images are output with binary pixel values indicating whether a pixel is in the dense phase (red), dilute phase (green), or background (blue).Delete default filenames from parse_args().Add makeimagestack.py to convert output from ImageXpress Confocal microscope to ImageJ compatible color and Z-stacks. Tested on Z-series output from ImageXpress as well as single-Z data.Implement handing mark-and-find acquisitions, including modifications to get_leica_xml_info() and file output handling, fix bit-depth info in ImageJ metadata, reformat output .csv file.Fix problem with writing condensate_masks_color images for 2D images.Remove default from args.input. Tested on 46 images and runs and outputs all files.Turn off low contrast warnings in io.imsave()Remove make_ij_lut import.Add combinechannels.py. Tested on Sandhya's images and runs correctly. rescaletif.py still not fully implemented.\Add makemontage.py. Tested, but not exhaustively.Fix artifact in writing condensate_masks_color image where the corner of the droplet bounding box overlays green on top of red or blue pixels by zeroing green value in all pixels where red or blue value is greater than zero.Implement writing a diagnostic image where each grayscale image at the dense-phase z-plane is shown with a color overlay showing dense phase in red and dilute phase in blue, change dilutemaskinnerradius default to 2 pixels, implement asking user whether to overwrite existing directories.Fix error with nblobsigmas command line argument parsing.Fix very stupid error regarding indentation of io.imsave() calls relative to the loops they are nested in. Script runs much faster now. Implement writing an output .csv with image acquisition parameters. v0.2Finish implenting lif2tif.py. Runs and gives output for all image types tested, and appears to give identical results for xyz 3-color tilescan images as compared with Fiji bioformats importer. v0.1.Show default values with --help argument, make several parameters configurable on the command line, make rolling ball background subtraction optional and turned off by default, change distance transform threshold argument to an absolute number rather than a percentage threshold, implement returning mean intensity over the entire droplet. Start work on .lif file processor. v0.2Implement function to create a segmented image with droplet index text superimposed.Fix bug dealing with images with no droplets. Add droplet size filter.Fully functional! v0.1Finish implementation of all functions required for basic segmentation, dense/dilute phase intensity measurement, dense phase volume measurement, and saving data. Runs without errors, but seems to be giving spuriously high dense phase intensities in the red channel.