Skip to content. | Skip to navigation

Sections
Personal tools
You are here: Home Forum Using data sets vim-1: high resolution imagery

vim-1: high resolution imagery

Up to Using data sets

vim-1: high resolution imagery

Posted by Aswin Sankaranarayanan at April 05. 2012

I am trying to get a hang of the encoding model described in Kay et al. Identifying natural images from human brain activity, Nature 2008.

I am trying to go over the steps described in supplementary methods part 5.

The method to localize the RF associated with a voxel requires the stimuli at multiple resolutions. However, the data in stimuli.mat is already at a lower resolution (128x128).

 

any chance the original stimuli be shared as well.

 

regards,

aswin

Re: vim-1: high resolution imagery

Posted by Mark Lescroart at April 07. 2012

The full-resolution stimuli will be shared soon. 

cheers,

Mark

Re: vim-1: high resolution imagery

Posted by Jeff Teeters at April 07. 2012

The full-resolution stimuli are now available in the downloads folder for the data set.  The readme file has been updated describing the additional files.
Thanks Mark!

Re: vim-1: high resolution imagery

Posted by Aswin Sankaranarayanan at April 14. 2012

Thanks a lot for sharing the full-resolution stimuli

 

just wanted to quickly point out that Stimuli_Val_FullRes.mat is missing in the downloads folder

Re: vim-1: high resolution imagery

Posted by admin at April 14. 2012

Sorry.  It's there now.  Thanks for letting us know.

 

Re: vim-1: high resolution imagery

Posted by Aswin Sankaranarayanan at April 14. 2012

thanks once again for the fast responses!!

Re: vim-1: high resolution imagery

Posted by Nicholas Hunter at October 30. 2012

Do you happen to have the full-resolution images without the "fovea blackout frame"? I would like to have these so certain types of image processing techniques don't alias the frame.

Re: vim-1: high resolution imagery

Posted by Mark Lescroart at November 02. 2012

Hi Nicholas, 

The stimuli we provided are exactly the stimuli that the subjects saw in the experiment, and were provided for the purpose of modeling our data. If you're interested in a set of natural images to use for other purposes, there are many other sources of them on the web. If you're interested in modeling our data, then the images we provided would seem to be the most relevant images; it doesn't seem like a good idea to model responses to image features in the corners of the image that the subjects did not see, or to remove a feature (the circular boundary) that was actually there in the experiment. Or, perhaps I should say, it doesn't seem to be a good enough idea to justify the nuisance of slogging through old file directories...

Cheers,

Mark

Powered by Ploneboard
Document Actions