This is the code used to produce the results and visualizations published in

Schmittwilken, L., Wichmann, F. A., & Maertens, M. (2023). Standard models of spatial vision mispredict edge sensitivity at low spatial frequencies. Vision Research, 222. doi:10.1016/j.visres.2024.108450

Setup

Install all the libraries in requirements.txt.

pip install -r requirements.txt

Note: we have used an older version of python-psignifit here, which is not available anymore. Therefore, we decided to add it to the repo directly in the folder psignifit. You can find information on the newest version of psignifit here.

Description

The repository contains the following:

• Code for empirically testing edge sensitivity in noise and the psychophyical data: experiment. If you want to run the experiment, you need to install the HRL library. For this, follow the instructions here.

• Code to set up and optimize all the variations of the standard spatial vision model as described in the paper: simulations. To create the noise masks for the simulation, run create_noises.py. To optimize the single-scale model, run optimize_single.py. To optimize the multi-scale model(s), run optimize_multi.py. Since all variable parameters are part of the normalization-step, both scripts will first run and save all model outputs to disc to reduce compute time.

• Code to create the visualizations from the manuscript and explore the empirical data and the model(s): visualizations. In order to re-create the deviance plots and model-psychometric-curves, you first need to run the simulations to produce the respective results.

• An old version of python-psignifit: psignifit

Authors and acknowledgment

Code written by Lynn Schmittwilken ([email protected])