Code designed for analysis of tongue contour data - produces three metrics (Procrustes analysis, Modified Curvature Index and Fourier analysis)
Switch branches/tags
Nothing to show
Clone or download
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
test_data added test data Jun 28, 2015
LICENSE Initial commit Jun 28, 2015
README.md added test data Jun 28, 2015
shape_analysis.py Improvements to MCI algorithm robustness Aug 7, 2017

README.md

tshape_analysis

This script contains code for several shape analysis metrics designed to be used on contours taken from ultrasound images of the tongue.

The analysis metrics are:

  • Procrustes analysis
  • Modified Curvature Index
  • Fourier analysis of the tongue profile

The code is written in Python.

Data type and output file

  • The data this script is designed to run on consists of .csv files containing x and y coordinate points for a series of tongue shape contours. The filename/s of the .csv files should be structured as: identifier_symbolforsegmenttype.csv (e.g. ‘01FC_ara’ where 01FC is the participant, and ‘ara’ is the stimulus item or segment type).

  • The script assumes the first column of data in the .csv is the x coordinates for the first shape, and the second column is the y coordinates for the first shape. Subsequent columns in the same .csv file are for additional shapes.

  • If your .csv files have header lines, the n_header_lines parameter (by default set to 0) can be used to ignore the header data in your files.

  • To run the Procrustes analysis, there must be a file in the directory containing the resting, or baseline comparison shape. The name of this file must be structured as: identifier_rest.csv. If this file is not available, no Procrustes analysis will be performed.

  • The default output filename is: shape_analysis_data_out.csv. The output file contains a row for every contour, and includes the following data:

    • ID (taken from filename)
    • symbol (taken from filename)
    • repetition (based on location in .csv file)
    • MCI - Modified Curvature Index value
    • procrustes - Procrustes Analysis value
    • real_1 - real part of 1st Fourier coefficient
    • imag_1 - imaginary part of 1st Fourier coefficient
    • mod_1 - modulus (magnitude) of 1st Fourier coefficient
    • real_2 - real part of 2nd Fourier coefficient
    • imag_2 - imaginary part of 2nd Fourier coefficient
    • mod_2 - modulus (magnitude) of 2nd Fourier coefficient
    • real_3 - real part of 3rd Fourier coefficient
    • imag_3 - imaginary part of 3rd Fourier coefficient
    • mod_3 - modulus (magnitude) of 3rd Fourier coefficient

Instructions for use

  • The script expects the .csv files containing the data to be in the current working directory.
  • The script can be run in Python or in a terminal window
  • If you use the terminal window, type in the file path (including the filename) for ‘shape_analysis.py’. Press Enter to execute the script.
  • If you are in Python, type: import shape_analysis then shape_analysis.doIt()

Test data

Contained in the ‘test data’ folder of this repository are a number of .csv files with tongue contours for various shapes. The filenames are constructed as required by the script - i.e. ’01FC’ is the participant ID, and ‘ara’ etc. is the utterance. For all of these files, the tongue contours represent the midpoint of the middle segment of the utterance, i.e. /r/, /l/, /g/ etc. There are 6 repetitions of each segment type. The data is in the mid-sagittal plane, collected from a typical, female speaker of American English.

The authors make no promise of continued support or development of the code contained in this repository.