*************README file for********************************************************************
Frequency analysis of gravure printed finger patterns using fast Fourier transform (MATLAB code)
************************************************************************************************
Last modified: 2024-03-15 (yyyy-mm-dd)

This dataset was generated by Pauline Rothmann-Brumm (2023) as part of her dissertation at the Technical University of Darmstadt, Germany.

Title of the dissertation: Visualisierung, Analyse und Modellierung von fluiddynamischen Musterbildungsphänomenen im Zylinderspalt unter Anwendung von Maschinellem Lernen (German) / Visualization, analysis and modeling of fluid dynamic pattern formation phenomena in the cylinder gap using machine learning (English translation)

---------------------

DATASET DESCRIPTION

URL: https://tudatalib.ulb.tu-darmstadt.de/handle/tudatalib/3839

This dataset contains MATLAB code for frequency analysis of gravure printed finger patterns from the HYPA-p dataset (see https://tudatalib.ulb.tu-darmstadt.de/handle/tudatalib/3841). The developed algorithm performs 1-dimensional fast Fourier transform (FFT) on selected image data from the HYPA-p dataset, i.e. on all 'L-fields'. The peaks of the FFT-spectra correspond to the dominant distances of the fingers. Peak finding is performed after smoothing of the FFT-spectra with pre-defined filter kernels.

---------------------

MATLAB CODE

Step by step guide:

1. Download folder 'code_FFT_ScalingFingering.zip' from https://tudatalib.ulb.tu-darmstadt.de/handle/tudatalib/3839. Unzip folder.

2. Optionally: Download folders '3_Temporary.zip' and '4_Output.zip' from https://tudatalib.ulb.tu-darmstadt.de/handle/tudatalib/3839. Unzip the folders and copy them to the folder 'code_FFT_ScalingFingering'. '3_Temporary' contains intermediate results, see 5. '4_Output' contains all the results of the frequency analysis from the dissertation of Pauline Rothmann-Brumm (2023). These results can be recreated with the MATLAB code (using 'preproc' = 'gray' as user input in 'main.m'. This means that the input images were converted to grayscale before performing the FFT.)

3. Requirements:
- MATLAB R2022b with 'Signal Processing Toolbox', 'Image Processing Toolbox' and 'Statistics and Machine Learning Toolbox'
- matlab2tikz (http://www.mathworks.com/matlabcentral/fileexchange/22022 or https://github.com/matlab2tikz/matlab2tikz)
- RobustSP toolbox (https://github.com/RobustSP/toolbox) --> Functions: MlocHUB.m, MlocTUK.m, whub.m, wtuk.m, madn.m
Copy matlab2tikz and RobustSP toolbox to folder '2_Ressources' --> '1_Functions'.

4. Open 'main.m' in the folder 'code_FFT_ScalingFingering'. 

5. To reproduce the results from the dissertation of Pauline Rothmann-Brumm (2023), you can either run 'main.m' completely, with all of the printed finger patterns from the dissertation as input (single files, no subfolders) - this will take a long time! - or you can load one of the .mat-files from the folder '3_Temporary' (contains intermediate results), skip some cells of 'main.m' and only run selected cells (using the 'run cell' functionality of Matlab). E.g. you can run lines 1-61, skip lines 62-67, then load '2023_06_25_202602_data_smooth_pks_gr.mat' into the workspace and thus skip the time-consuming FFT (lines 68-71) as well as the smoothing (lines 72-75), peak finding (lines 76-80) and grouping of the data (lines 81-84) and continue from line 85. The mentioned printed finger patterns ('L-fields') can be downloaded from the HYPA-p dataset (https://tudatalib.ulb.tu-darmstadt.de/handle/tudatalib/3841). In total, 3,904 L-fields with a total size of 2 TB are available for download. However, to test selected functionalities of 'main.m' (FFT, smoothing and peak-finding), one exemplary L-field is enough. Copy it to the folder '1_Input'.

6. Either run 'main.m' completely or only run selected cells (using the 'run cell' functionality from Matlab) after loading a .mat-file from the folder '3_Temporary' as described in 5.

---------------------

CONTACT

Pauline Rothmann-Brumm
Technical University of Darmstadt, Department of Mechanical Engineering, Institute of Printing Science and Technology (IDD), Magdalenenstr. 2, 64289 Darmstadt, Germany
rothmann-brumm@idd.tu-darmstadt.de