/SIFLIM

Software for image selection and marker labeling for FLIM using feature space projections

Primary LanguagePython

SIFLIM

Software for image selection and marker labeling for FLIM using feature space projections

Requirements

Create virtual environment with Python 3.8 (ex. anaconda or miniconda)

  • pip install -r requirements.txt

Generanting Dataset of Patches from Input Images

SIFLIM can recieve as input an OPFDataset with each sample being an NxNx3 patch extracted from a given image's superpixel segmentation.
To create such dataset you can run
python generate_patches_dataset.py <INPUT> <OUTPUT.zip> <PATCHSIZE> <MAX-SAMPLES-PER-CLASS>
on your virtual environment, where

  • <INPUT> is the path to the fileset of original .png images
  • <OUTPUT.zip> is the desired output file path/name
  • <PATCHSIZE> is the patch size (e.g. = 75 generates 75x75x3 patches)
  • <MAX-SAMPLES-PER-CLASS> is the max number of generates sample points per image