Skip to content

Latest commit

 

History

History
58 lines (45 loc) · 2.5 KB

README.md

File metadata and controls

58 lines (45 loc) · 2.5 KB
Raw

EIT Image Reconstruction Algorithm

This is a submission for the Kuopio Tomography Challenge.

Authors

  • Amal Mohammed A Alghamdi (DTU), Denmark
  • Martin Sæbye Carøe (DTU), Denmark
  • Jasper Marijn Everink (DTU), Denmark
  • Jakob Sauer Jørgensen (DTU), Denmark
  • Kim Knudsen (DTU), Denmark
  • Jakob Tore Kammeyer Nielsen (DTU), Denmark
  • Aksel Kaastrup Rasmussen (DTU), Denmark
  • Rasmus Kleist Hørlyck Sørensen (DTU), Denmark
  • Chao Zhang (DTU), Denmark

Addresses

DTU: Technical University of Denmark, Department of Applied Mathematics and Computer Science Richard Petersens Plads Building 324 2800 Kgs. Lyngby Denmark

Description of the algorithm

We have used the provided code for the EIT image reconstruction with the following modifications:

  • The Otsu segmentation algorithm has been replaced by the Chan-Vese segmentation algorithm from scikit-image.
  • Additional generalized Tikhonov regularization has been added to penalize more when close to the missing electrodes (and boundary). The regularization matrix is a diagonal matrix. For example, for difficulty level 5, the amount of penalty added to different regions of the image, can be seen in the image below:

Installation instructions

To run our EIT image reconstruction algorithm, you will need:

  • Python 3.x
  • Required Python libraries (listed in requirements.txt)
  • Access to the provided dataset (not included in this repository)

Usage instructions

python main.py path/to/input/files path/to/output/files difficulty

Examples

Phantom Ref Level 1 Level 4 Level 7
a
b
c
d

Scores for each phantom and difficulty 1,4 and 7:

Phantom Level 1 Level 4 Level 7
a 0.835 0.737 0.526
b 0.761 0.563 0.499
c 0.942 0.918 0.844
d 0.767 0.774 0.751

Scores have been computed using our own implementation of the scoring function based on scikit learn.

License

All files in the repository come with the Apache-v2.0 license unless differently specified.