Collaborative development of models of early visual cortex.
Good models. Good datasets. Data-driven development of better models.
We built this site because we want to support collaboration across
lab boundaries, disciplinary boundaries and coding styles.
We also want to support direct comparison between different approaches.
We all have a sense of what "the standard V1 model" is, but we don't
always agree on the details. Exactly what does the log-polar transform look
like? Exactly how should divisive normalization be implemented? How close
is the Gabor approximation, after all? Virtual V1sion aims to provide a sand box in which we can play with
these questions. Interrogation of modeling results is supported by
interactive visualizations and a handy "Export .csv" button.
Get acquainted with existing models on the Model page. If you
have a better model, visit the Repo page to discover how to coontribute
it to the collection and take advantage of the sand box and visualization
tools hosted here.