The goal of team-rhodes-p3 is to detect neurons in images (in the format of time-series). This project is intended to be the solution for project 3, course CSCI8360 at the University of Georgia.
This project requires Python 3.x with cv2, imageio, thunder-extraction, joblib, and image_slicer libraries installed.
The dataset consists of a set of 28 samples (19 training samples and 9 test samples). Eeach sample is a a set of .tiff images grouped into a directory.
Since the size of the dataset is large, the data is not included in the repository, however a script to download the dataset is provided in the repository. To download the data navigate to scripts/ directory and run the following in command line (note that google-cloud-sdk is required for the script):
$ ./get_files.sh
You may download the source code and simply run the following command:
$ python3 NMF.py -i 'path/to/images/'
List of command line arguments to pass to the program are as follows:
--input: Path to image sets.
--output: Path to save json files to.
--n_jobs: number of jobs to spawn in parallel
--save_individual: if set each dataset will be saved separately.
--n_comps: number of components to estimate per block.
--iters: max number of algorithm iterations.
--perc: the value for thresholding (higher is more thresholding).
--chunk_size: width and height of chunk, two values.
--overlap: value determining whether to merge.
--custom_config: if true use the best configuration.
The see the above list in command line execute the following command:
$ python3 NMF.py -h
Alternatively, you can install the package using pip as follows:
$ pip install --user -i https://test.pypi.org/simple/rhodes
In this case you can use the other functionality of this package to augment and prepare the dataset for Tiramisu model that can be obtained from here:
The code in this repository is free software: you can redistribute it and/or modify it under the terms of the MIT lisense.
For questions please email one of the authors: