This is the code used to produce the results published in

Betz T, Shapley R, Wichmann F A, Maertens M (2015) Noise masking of White's illusion exposes the weakness of current spatial filtering models of lightness perception. Journal of Vision.

The repository contains the following:

-Matlab code for generating noise stimuli, kindly provided by Dr. Salmela, and adapted to fit the needs of the present work: stimuli

-The code for running the experiment (experiment.py; requires hrl to run). Due to the file size, the noise masks themselves are not provided, but can be created with generate_noisemasks.m

-The data from the psychophysical experiment: exp_data

-A script to analyze the experimental data and generate figures: analyze_noise_data.py

-Scripts to analyze the different lightness models with noise:

• evaluate_models.py to compute model responses for ODOG and the Dakin-Bex model, and to create result figures for all models
• evaluate_matlab_models.m to compute responses to noise stimuli for BIWAM and FLODOG
• code for the BIWaM model (CIWaM), kindly provided by Dr. Otazu
• code for the FLODOG model (f_l_odog_models), kindly provided by Dr. Robinson
• implementations of the ODOG and Dakin-Bex model are located in a separate repository lightness_models

-The data from the model simulations: data

-A script to create the phase shuffled version of the COBC illusion (Figure 12 in the article): create_dakin_bex_demo.py

Some of the python scripts require ocupy to run.

Documentation for the analysis scripts is sketchy, but the models themselves are well documented. Get in touch with Torsten in case you have questions.