Learning orientation-invariant representations enables accurate and robust morphologic profiling of cells and organelles
Authors/Creators
- 1. Stanford University
- 2. University of North Carolina at Chapel Hill
Description
Datasets supporting the publication, "Learning orientation-invariant representations enables accurate and robust morphologic profiling of cells and organelles".
Dataset 1, PCST: Profiling Cell Shape and Texture (PCST) benchmark
Synthetic dataset created for this publication . See `README_PCST.md` for more details.
File `PCST.tar.gz` is the full (very large) dataset. Each image is one file.
File `PCST_subsets_processed.tar.gz` is a subset and processed. Each file here is the first 15,000 images in tensors: one tensor is the 3-channel image, and the second tensor is the segmentation mask.
On a unix system, open data with `tar -xzvf PCST_subsets_processed.tar.gz` for subset data, or `tar -xzvf PCST.tar.gz` for the full dataset.
Dataset 2, mefs: Fluorescence Image of Mouse Embryonic Fibroblast seeded on Circle and Triangle micro-patterns.
The original dataset is from the paper "A robust unsupervised machine-learning method to quantify the morphological heterogeneity of cells and nuclei". (See the paper's data availability statement for the original dataset). The dataset here is the segmentations and our CellProfiler segmentation pipeline. See README.md for more details.
On a unix system, open data with `tar -xzvf PCST.tar.gz`'