Colour in Context
Research group
Computer Vision Center

The Brightness Induction Wavelet Model (BiWaM) online

BiWaM is a wavelet-based computational model that attempts to reproduce the perceptual processes involved in visual brightness induction (Otazu X, Vanrell M, Parraga CA, "Multiresolution wavelet framework models brightness induction effects", Vision Research, 48(5),733-751, 2008).

BIWaM takes an input image and considers it to be the stimulus received by the observer. The model then generates an output image that is an estimation of the image as "perceived" by the same observer (perceptual image). We would like to stress that, since this output image corresponds to the perceived image, the input and output images look very different. You will not see two identical images: YOUR visual system is processing the output image for a second time while you watch it onscreen. As a result, all the useful information that the output image provides is contained in its numerical values (LMS* or sRGB values), rather than in the actual appearance of the image.

The user simply has to upload a colour image in any common image file format and specify the colour space of the image (LMS* or sRGB). The system will output the perceptual image in the same colour space.

*Stockman and Sharpe (2000) 2-deg cone sensitivities

Select file to upload
Max file size = 5 MB

Color space:  LMS sRGB          Maximum contrast factor:   (default value 1.000 , rang value [0.00 -2.00])

Select the option to fill the parameters

The BIWaM only needs to know the physical dimensions of the scene and the observer's position to work. By choosing the first option, the BIWaM will regard your monitor as the ACTUAL monitor you were using to look at the image fullscreen. Introduce the distance from which you are looking at the screen (in cm) and the monitor diagonal size in inches.
The second option allows a more technical measure (angular image width in degrees) to be entered.

Option 1

Distance of monitor (cm):      
Monitor diagonal size (inch):

Option 2

Image width visual angle (degrees)

Note : This program accepts many images formats (.gif, .jpg, .jpeg, .png, .tif, .bmp, .ppm) but we recommend uncompressed image formats (.tif, .png, .ppm) because the other formats may have suffered spatiochromatic manipulations that render them unsuitable for this kind of modelling. The model's output is in .png format in all cases.