Research data and code supporting "Rideaux & Welchman (2021) Exploring and explaining properties of motion processing in biological brains using a neural network - Journal of Vision"
Repository URI
Repository DOI
Type
Dataset
Change log
Authors
Rideaux, Reuben
Welchman, Andrew https://orcid.org/0000-0002-7559-3299
Description
This repository contains the data and code presented in "Rideaux & Welchman (2021) Exploring and explaining properties of motion processing in biological brains using a neural network - Journal of Vision". Data includes the 10 trained MotionNet_xy neural networks presented in the manuscript and example images used to train MotionNet_xy. Code includes scripts used to train and test MotionNet_xy. See the readme.txt file contained in the data_and_code folder for more information.
Version
Software / Usage instructions
MATLAB PsychToolBox Psignifit Python TensorFlow Numpy Pickle Gzip Scipy MatPlotLib PsychoPy PyGame See the readme.txt file contained in the data_and_code folder for more information.
$$ \ $$
Opening the 'training_image_sequences.1' and 'training_image_sequences.2' files requires the python packages "Gzip" and "Pickle". These files will be read when the "train_network.py" script is run; however, if you wish to inspect them you can use the following commands from a python terminal:
$$ \ $$
### beginning of commands
import gzip, pickle
infile = gzip.GzipFile('training_image_sequence.1', 'rb')
obj = infile.read()
data = pickle.loads(obj)
infile.close()
### end of commands
$$ \ $$
These commands will unzip and unpickle the file into the structure "data". Note, it may take a couple of minutes to unzip the files depending on your machine.
Keywords
Machine Learning