LimSim 1.0

An example of lateral inhibition in the eye of the horseshoe crab

Please wait for the applet to initialize...

What you see when the applet first loads is a sample low contrast image consisting of a 20 x 20 grid of pixels, each representing a real (float) value that has been mapped to 8 bit greyscale. When you hit "OK," the simulation is run and the result of the lateral inhibition is displayed. With the initial settings you should find that the "A" (or whatever character), while a bit ligher itself, is surrounded by very light cells -- thereby enhancing the contrast, making the "A" easier to read.

How to use:

Load Presets: Allows you to load one of a number of preset characters against varying contrast backgrounds or positions.

Set Parameters: Brings up a window that allows you to tweak a number of parameters in the simulation.

Strength: Inhibitory strength, same as the value set in the Strength field of the control panel.

Length: Length Constant, same as the value set in the Length Constant field of the control panel.

OK / Reset: Click to begin the simulation.

Want to try this as an application? You'll need to have a Java Runtime installed (I recommend the JRE from Sun). Then download and unzip this zip file: LimSim.zip. Sorry, it's only available as a zip for now...

Parameters:

Strength (float): The strength of the inhibition. A low value results in low inhibition, while a high value will result in very large inhibition.
Length Constant (int): The integral distance of the reach of the inhibition. The larger the value, the farther away each cell will have its effects felt.
Upper/Lower Limit (float): Range of values possible in each cell. Note that the display will only show values 0 to 1, inclusive.
Epsilon (float): Currently unused.
Normalize: If checked, the strongest signal in the matrix will be normalized to 1, and all other values will be likewise scaled. The effects of inhibition will be even more pronounced if you set this.
Multiple Iterations: If checked, the simulation will run through multiple iterations to more closely approximate the effects of inhibition. Without, a single pass is made over the data.
Iterations: How many iterations you wish to run if Multiple Iterations is checked.