Skip to main content

Multidimensional scanpath comparison

Project description

Build Status codecov Documentation PyPIversion License: MIT Build status DOI

multimatch-gaze

Reimplementation of MultiMatch toolbox (Dewhurst et al., 2012) in Python.

The MultiMatch method proposed by Jarodzka, Holmqvist and Nyström (2010), implemented in Matlab as the MultiMatch toolbox and validated by Dewhurst and colleagues (2012) is a vector-based, multi-dimensional approach to compute scanpath similarity.

For a complete overview of this software, please take a look at the Documentation

The method represents scanpaths as geometrical vectors in a two-dimensional space: Any scanpath is build up of a vector sequence in which the vectors represent saccades, and the start and end position of saccade vectors represent fixations. Two such sequences (which can differ in length) are compared on the five dimensions ‘vector shape’, ‘vector length’ (saccadic amplitude), ‘vector position’, ‘vector direction’ and ‘fixation duration’ for a multidimensional similarity evaluation (all in range [0, 1] with 0 denoting maximal dissimilarity and 1 denoting identical scanpaths on the given measure). The original Matlab toolbox was kindly provided via email by Dr. Richard Dewhurst and the method was ported into Python with the intent of providing an open source alternative to the matlab toolbox.

Installation instructions

It is recommended to use a dedicated virtualenv:

# create and enter a new virtual environment (optional)
virtualenv --python=python3 ~/env/multimatch
. ~/env/multimatch/bin/activate

multimatch-gaze can be installed via pip. To automatically install multimatch-gaze with all dependencies, use:

# install from pyPi
pip install multimatch-gaze

Support/Contributing

Bug reports, feedback, or any other contribution are always appreciated. To report a bug, request a feature, or ask a question, please open an issue. Pull requests are always welcome.

Examplary usage of multimatch-gaze in a terminal

required inputs: - two tab-separated files with nx3 fixation vectors (x coordinate in px, y coordinate in px, duration) - screensize in px (x dimension, y dimension)

multimatch data/fixvectors/segment_10_sub-19.tsv data/fixvectors/segment_10_sub-01.tsv 1280 720

optional inputs:

if scanpath simplification should be performed, please specify in addition - –amplitude-threshold (-am) in px - –direction-threshold (-di) in degree - –duration-threshold (-du) in seconds

Example usage with grouping:

multimatch data/fixvectors/segment_10_sub-19.tsv data/fixvectors/segment_10_sub-01.tsv 1280 720 --direction-threshold 45.0 --duration-threshold 0.3 --amplitude-threshold 147.0

REMoDNaV helper:

Eye movement event detection results produced by REMoDNaV can be read in natively by multimatch-gaze. To indicate that datafiles are REMoDNaV outputs, supply the --remodnav parameter.

multimatch data/remodnav_samples/sub-01_task-movie_run-1_events.tsv data/remodnav_samples/sub-01_task-movie_run-2_events.tsv 1280 720 --remodnav

REMoDNaV can classify smooth pursuit movements. As a consequence, when using REMoDNaV output, users need to indicate how these events should be treated. By default, multimatch-gaze will discard pursuits. In some circumstances, however, it can be useful to include pursuit information. Moving stimuli for example would evoke a pursuit movement during visual intake. When specifying the --pursuit keep parameter, the start and end points of pursuits will be included in the scanpath.

multimatch data/remodnav_samples/sub-01_task-movie_run-1_events.tsv data/remodnav_samples/sub-01_task-movie_run-2_events.tsv 1280 720 --remodnav --pursuit keep

References:

Dewhurst, R., Nyström, M., Jarodzka, H., Foulsham, T., Johansson, R. & Holmqvist, K. (2012). It depends on how you look at it: scanpath comparison in multiple dimensions with MultiMatch, a vector-based approach. Behaviour Research Methods, 44(4), 1079-1100. doi: 10.3758/s13428-012-0212-2.

Dijkstra, E. W. (1959). A note on two problems in connexion withgraphs. Numerische Mathematik, 1, 269–271. https://doi.org/10.1007/BF01386390

Jarodzka, H., Holmqvist, K., & Nyström, M. (eds.) (2010). A vector-based, multidimensional scanpath similarity measure. In Proceedings of the 2010 symposium on eye-tracking research & applications (pp. 211-218). ACM. doi: 10.1145/1743666.1743718

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

multimatch_gaze-0.1.0.tar.gz (24.8 kB view details)

Uploaded Source

Built Distribution

multimatch_gaze-0.1.0-py3-none-any.whl (42.3 kB view details)

Uploaded Python 3

File details

Details for the file multimatch_gaze-0.1.0.tar.gz.

File metadata

  • Download URL: multimatch_gaze-0.1.0.tar.gz
  • Upload date:
  • Size: 24.8 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/1.13.0 pkginfo/1.5.0.1 requests/2.21.0 setuptools/40.8.0 requests-toolbelt/0.9.1 tqdm/4.31.1 CPython/2.7.16

File hashes

Hashes for multimatch_gaze-0.1.0.tar.gz
Algorithm Hash digest
SHA256 ea624b78c116dcda850d7b2011e49ad78033e261b9002f97acc7ad3cef33b7bf
MD5 9e54de2968c3af094c336ae96b1934df
BLAKE2b-256 a187bdb2c6fb3dc0dd5f56f30db89ceab3fb107880e3524d91d9364f8b0f0523

See more details on using hashes here.

File details

Details for the file multimatch_gaze-0.1.0-py3-none-any.whl.

File metadata

  • Download URL: multimatch_gaze-0.1.0-py3-none-any.whl
  • Upload date:
  • Size: 42.3 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/1.13.0 pkginfo/1.5.0.1 requests/2.21.0 setuptools/40.8.0 requests-toolbelt/0.9.1 tqdm/4.31.1 CPython/2.7.16

File hashes

Hashes for multimatch_gaze-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 02b06f1765e8a34c5aed95c6128400df90f9d9b1b77487b6f308b53724e4c9a0
MD5 911f80fa6f40b986dc563642224fc666
BLAKE2b-256 2fba7b25256f945adc19082562aeab903d23446467259ee33b0b71469e679bf8

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page