s_HumanWavelengthDiscrimination

We discriminate two lights with different wavelengths by comparing the pattern of cones absorptions generated by the lights. The separation in the pattern of cone response depends, in turn, on the photon noise in the absorptions.

This script calculates the cone absorptions for lights at different mean luminance levels. The graph shows how the absorption data generated by the two wavelengths become better separated, as the mean illumination level grows.

This script also shows some of the details of how to create the properties of a human cone sensor array.

Create a sample human optics and cone mosaic sensor
Compute cone absorption as a function of mean signal level

s_initISET;
rng('default')

Create a sample human optics and cone mosaic sensor

Standard optics and human retina simulations can be created this way. You can also change the parameters from the standard, say for simulating the periphery or for simulating people with biological variability.

% Human optics
oi = oiCreate('human');

% Create a typical cone mosaic and show a little picture
cSensor = sensorCreate('human');
sensorConePlot(cSensor);

Compute cone absorption as a function of mean signal level

We will image uniform fields of monochrome light on a human sensor. Then get the photon absorptions in a 100 ms flash. THen we plot the absorptions as a 3D graph in the next cell.

% These are the two wavelengths and the list of scene luminance levels
wSamples = [520  530];
nWave = length(wSamples);
luminance = [10 50 200];
nLevels = length(luminance);
sceneSize = 128;

% We will extract the cone absorptions for plotting into these variables
L = cell(1,length(wSamples));
M = cell(1,length(wSamples));
S = cell(1,length(wSamples));
% The sensor color filters are black (K, a blank spot), L, M and S
% The default human sensor is 0,6,3,1 for K,L,M,S
slot = [2 3 4];   % L,M,S positions in the sensor

% Make a series of scenes at different wavelengths and peak readiances.
% Then, compute the sensor response.
scene  = cell(1,nWave);
sensor = cell(1,nWave);

vcNewGraphWin([],'tall');
for rr = 1:nLevels
    subplot(nLevels,1,rr);
    for ww=1:length(wSamples)

        % Create a monochromatic scene and set the radiance
        % The wavelength is specified in wSamples.
        scene{ww} = sceneCreate('uniform monochromatic',wSamples(ww),sceneSize);
        % scene{ww} = sceneSet(scene{ww},'peak photon radiance',peakRadiance(rr));
        scene{ww} = sceneAdjustLuminance(scene{ww},luminance(rr));

        % Compute the spectral irradiance at the retina
        oi = oiCompute(scene{ww},oi);

        % Create a human sensor that will integrate for 100 ms
        sensor{ww} = sensorCreate('human');
        sensor{ww} = sensorSet(sensor{ww},'exposure time',0.10);

        % Compute the sensor absorptions
        sensor{ww} = sensorCompute(sensor{ww},oi);
        sensor{ww} = sensorSet(sensor{ww},'name',sprintf('wave %.0f',wSamples(ww)));

        % If you want to have a look at the image, run this line.
        % vcAddAndSelectObject(sensor{ww}); sensorWindow;
    end

    for ww=1:length(wSamples)
        L{ww} = sensorGet(sensor{ww},'electrons',slot(1));
        M{ww} = sensorGet(sensor{ww},'electrons',slot(2));
        S{ww} = sensorGet(sensor{ww},'electrons',slot(3));

        % For simplicity in plotting, make the absorptions same length
        n = min(100,length(L{ww}));
        S{ww} = S{ww}(1:n); M{ww} = M{ww}(1:n); L{ww} = L{ww}(1:n);
    end

    % Plot the absorptions
    sym = {'b.','g.','r.'};
    az = 65.5; el = 30;
    for ww=1:length(wSamples)
        title(sprintf('Luminance (cd/m^2): %.0f',luminance(rr)));
        s = mod(ww,length(sym))+1;
        plot3(L{ww}(:),M{ww}(:),S{ww}(:),sym{s})
        view([az el])
        hold on
    end
    xlabel('L-absorptions'); ylabel('M-Absorptions');
    zlabel('S-absorptions'); axis square; grid on

end

s_HumanWavelengthDiscrimination

Contents

Create a sample human optics and cone mosaic sensor

Compute cone absorption as a function of mean signal level