An open API service indexing awesome lists of open source software.

https://github.com/iitis/hsi_blood_detection


https://github.com/iitis/hsi_blood_detection

Last synced: 11 months ago
JSON representation

Awesome Lists containing this project

README

          

Description:
------------

Source code enabling replication of experiments in the paper
by M. Romaszewski, P.Glomb, M. Cholewa, A. Sochan
`**A Dataset for Evaluating Blood Detection in Hyperspectral Images**'
Journal article: https://doi.org/10.1016/j.forsciint.2021.110701

The dataset associated with that source code:
---------------------------------------------

The dataset is available online:
https://zenodo.org/record/3984905

Implementation:
---------------

Experiments were implemented in Python 3.6.9 using libraries:
numpy 1.16.4, scipy 1.3.1, scikit-learn 0.22.1, matplotlib 3.2.2

Usage:
------


  • Ensure that the dataset patch in ds_load.py is correct

  • Run experiments first, before generating results (unless it is dataset presentation)

Files:


  • ds_load.py - data loading and utility functions

  • results_util.py - implements detection performance measures

  • saveload.py - universal serialisation functions

  • target_detectors.py - MF detector implementation

  • two_stage_detector.py - implementation of the TSMF from the paper

  • experiment_classification.py - performs simple classification experiment

  • experiment_detection.py - performs detection experiments

  • fig_dataset.py - figures with dataset presentsion

  • fig_detection.py - figures with detection results

  • fig_example.py - detection examle with more detailed presentation

  • fig_means.py - presentation of spectra before/after applying TSMF

  • fig_no_pixels.py - figures visualising impact of the no. pixels on AUC

Warning:
experiments use the external blood library described in:
**Majda, Alicja, et al. "Hyperspectral imaging and multivariate analysis in the dried blood spots investigations" Applied Physics A 124.4 (2018): 312.**
Unless the library is provided, its spectra will replaced with a spectrum from the dataset.

License:
--------

The code is licensed under GNU General Public License v.3.0