Mathematica. Matlab. User-friendly Application in Matlab.

General Description

The model takes two luminance images. One is the standard or background image; the other is different in some way. The images are converted to contrast images, and then filtered by a contrast sensitivity function. The Minkowski distance between the two filtered images is normalized by the RMS contrast of the standard or background to give a d' estimate for the discriminability of the images. Setting the spatial integration exponent beta to 2 gives the visible contrast energy model which allows simpler code and faster execution ( Matlab version, R version).

Model applications

1) Detectability of image processing artifacts
2) Estimation target detection using simulated (exactly replicated) backgrounds

Capabilities and limitations

1) Images need not be square or have square pixels
2) Contrast is based on average luminance, so luminance in target region must be close to average luminance.
3) Masking is based on background RMS, so contrast energy in target region must be close to image average.
4) No provision for visual field inhomogeneity.
5) Linearization allows computation of threshold luminance target by multiplication of the difference of the luminance images by 1/d'.
6) Linearization requires that the contribution of the target to the contrast and masking calculations be negligible.
7) Contrast sensitivity function and maximum contrast sensitivity not valid for low average luminances (< 10 cd/m^2).
8) Contrast sensitivity function calculation is slow, but expository.

System requirements: Mathematica version tested on Mathematica 2.2 for SPARC. Matlab version tested on Matlab 5.2 for MacIntosh.

Data requirements

Input data

1) Number of rows and columns in each image
2) Pixels per degree of visual angle in row and column directions
3) Luminance values for each image

Data format and units

1) Images are assumed to be in Mathematica arrays.
2) Luminance units do not matter, they are converted to contrast.

Model output data

d' estimate

Uses of model output

1/d' times the difference luminance image is predicted signal threshold

References

Ahumada (1996 SID)
Simplified Vision Models for Image Quality Assessment

Ahumada and Beard (1996 SPIE)
Object Detection in a Noisy Scene Ahumada and Beard (1997 JOSA)
Image Discrimination Models Predict Detection in Fixed but not Random Noise

Ahumada and Beard (1997 SPIE)
Image Discrimination Models: Detection in Fixed and Random Noise

Rohaly, Ahumada, and Watson (1997 VR)
Object detection in natural backgrounds predicted by discrimination performance and models