Home

Welcome to the spatial-vision-model wiki!

This wiki gives instructions on how to use the code. For more details on how the model works and what data we can or cannot describe well with it, please refer to the journal article.

Install

The model is written in MATLAB and the basic model does not require any toolboxes. For some ways how we save images we require the image processing toolbox, however. Also, the distributed and parallel computing toolboxes are required to run on GPUs.

If you have MATLAB installed, download or clone the GitHub repository (red button at the upper right of the code page. Then simply add the folder to your path and you should be good to go.

Overview of the code

The whole model can be run with the early_vision_model functions, which primarily expect a luminance image in [cd/m^2] as an input and compute an internal representation with spatial frequency and orientation channel activities and an error term for those.

You can find more on the different possible settings and parameters for the whole function and which ones we used for which part of the article here.

You can read more on the parts of the model called by early_vision_model here.

There are a couple of support functions we provide with this toolbox for generating images, estimating thresholds for some simple experiments etc. You can get an overview of them here.

Finally, we provide a shortened version of the script we used to make the model illustration for our article as a demo showcasing the steps in the model as illustration.m .