Revision as of 14:05, 10 August 2015 by Rjpatruno (Talk | contribs)
Jump to: navigation, search


The simulations begin with an image (hyper-spectral or RGB for a simulated monitor).

We take the spectral image through a set of calculations from

  • radiance image to retinal irradiance
  • from retinal irradiance to L,M,S absorptions

The parameters (e.g., condition of the human optics, sampling distribution of the L,M,S cones, integration time) can be altered. These produce an image that we are currently storing in the rgc/images directory.

We then use the rgc software to simulate the responses of an array of retinal ganglion cells. The key scripts for doing this are in the scripts directory and start with 's_rgc*'. Two we are currently using are s_rgcSingleShapes.m and s_rgcSpikeFiles.m


The software is held in synapse/rgc.


We acquired a set of images from the virtual environment.

To simulate the cone absorptions we used ISET to create a spectral radiance image of the PNG file. We then passed the file through a simulation of the human optics to create an irradiance image.

- could go here -

We then created a simulated human cone mosaic.


We converted the irradiance image to cone photon absorption rates for a mosaic. These were stored in Matlab files rgcVE-Circle.mat and so forth

We then used the RGC simulations to produce a video of the spiking time series to this cone mosaic while the eye moved randomly across the image. The spikes were produced using the Pillow model.

The average of the spikes is after a 500 ms view is shown at the right.

Personal tools