University of Cambridge > Talks.cam > Computational Neuroscience > Computational Neuroscience Journal Club

Computational Neuroscience Journal Club

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Dr Jean-Pascal Pfister.

Statistical modeling of photographic images E P Simoncelli, Handbook of Image and Video Processing,pages 431—441. Academic Press, May 2005.

Over the last 15 years it has been increasingly popular to model the visual cortex as performing inference in a latent variable model for image statistics. The basic idea is that visual scenes consist of structural primatives, like local edges at particular orientations and scales. In any one image only a relatively small number of these primitives are assumed to be present, and so these models employ sparse latent variables that describe the probability of occurrence. Cortex is then modelled as representing the posterior distribution over these latent variables, given the retinal image.

In this journal club, I will review this modelling work, starting with independent component analysis, and sparse coding, which have been used to model simple cells. I will then describe their generalisation to Gaussian Scale Mixtures, which have been used to model both simple and complex cells. Finally I hope to describe the most recent work on these models by both Mike Lewicki and Eero Simoncelli. One of the main aims will be to assess this modelling work whilst wearing a machine learning hat.

This talk is part of the Computational Neuroscience series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

© 2006-2019 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity