Method and apparatus for determining visually perceptible differences between images

Claim 2

A method of comparing a first image signal and a second image signal, according to Claim 1, further comprising the step of: providing an image output having indications at locations thereon where a user is likely to perceive a difference between the images.

Claim 3

A method of comparing a first image signal and a second image signal, according to Claim 1, wherein the step of modifying the first and second image signals to model non-linear response includes adjusting a luminance level of the first and second image signals ; wherein preferably the step of modifying the first and second image signals to compensate for optical properties includes frequency filtering the first and second image signals ; and wherein preferably the step of modeling of the human visual system's ability to discern a particular feature in an image is a function of orientation and frequency content of the particular feature and a surrounding image.

Claim 4

A method of comparing a first image signal and a second image signal, according to Claim 2, wherein said image output is a free-field map having gray portions thereon indicating that a user is likely to perceive no difference between the images and having darker or lighter portions thereon indicating that a user is likely to perceive the second image as being darker or lighter, respectively, than the first image ; and wherein preferably said image output is an in-context map having a first and second color superimposed on locations of the reference image wherein the first color indicates that a user is likely to perceive the first image as darker than the second and the second color indicates that a user is likely to perceive the first image as lighter than the second.

Claim 5

Apparatus for determining visual differences between a first image signal and a second image signal, comprising: an amplitude non-linearity unit, for adjusting luminance of the first and second image signals; a contrast sensitivity unit, coupled to the output of said amplitude non-linearity unit, for determining visual sensitivity of the images as a function of frequency; a cortex unit, coupled to the output of said contrast sensitivity unit, for subdividing each of said first and second image signals into a plurality of cortex-filtered image signals which vary according to visual orientation and frequency content of image information; a contrast difference unit, coupled to the output of said cortex unit, for providing image signals indicative of the difference between said cortex-filtered first and second image signals at a plurality of visual orientation and frequency band combinations of said cortex filters; a threshold elevation unit, coupled to the output of said cortex unit, for determining the visual sensitivity of said images according to the frequency and orientation of information in the image signal; a threshold elevation comparison unit, coupled to the output of said threshold elevation unit, for determining an overall threshold according to the threshold associated with said first image and the sensitivity threshold associated with said second image; and a probability generator unit, coupled to the output of said contrast unit and the output of said threshold elevation comparison unit, for determining the probability that a user will visually perceive a difference between the first image and the second image.

Claim 6

Apparatus for determining visual differences between a first image signal and a second image signal, according to Claim 5, further comprising: a free-field map generator unit, for providing an image having gray pixels at locations wherein no perceptible differences between the first and second images are detected and having darker or lighter pixels at locations wherein the first image is predicted to be perceived as darker or lighter, respectively, than the second image , preferably an in-context map generator unit, for outputting the first image signal with a first and second color superimposed thereon at locations wherein the first image is predicted to be perceived as darker or lighter, respectively, than the second image, wherein preferably said cortex filter unit employs mesa filters defined by the function: where ρ represents frequency, tw represents transition width of a frequency band, and H represents a half-amplitude frequency , and wherein preferably said cortex filter unit employs fan filters defined by the function: where l is a designation of the particular orientation, φ is the angle of orientation, φ tw is the width of each of the fan filters, and φ c (l) is the midpoint angle of fan filter l.

Claim 7

Apparatus for determining visual differences between a first image signal and a second image signal, according to Claim 6, wherein said cortex filter unit employs a baseband filter defined by the function: where ρ is the frequency, H is 2 -(number of frequency bands) , and σ is a constant given by: wherein preferably said cortex filter unit retains radial frequencies greater than .5 cycles per pixel, and wherein preferably said threshold elevation unit is defined by the function: where Ix represents either the first image or the second image, i, j indicate a particular pixel, the value of b is 4.0, the quantity s equals 0.7 for the baseband image, 0.8 for images filtered through cortex filters having the two highest frequencies and 1.0 for all other images.

Claim 9

A method for determining visual differences between a first image signal and a second image signal, according to Claim 16, further comprising the step of: providing an image having gray pixels at locations wherein no perceptible differences between the first and second images are detected and having darker or lighter pixels at locations wherein the first image is predicted to be perceived as darker or lighter, respectively, than the second image , and preferably outputting the first image signal with a first and second color superimposed thereon at locations wherein the first image is predicted to be perceived as darker or lighter than the second image.

Claim 10

In a method of determining overall visual sensitivity as a function of radial frequency over a range of distances, the steps of: determining visual sensitivity as a function of radial frequency for a plurality of incremental distances within the range of distances; and deeming said overall visual sensitivity as a function of radial frequency equal to the maximum sensitivity value for each pixel ascertained during said determining step.

Claim 11

In a method of determining visual sensitivity as a function of image content, the step of low pass filtering a threshold elevation signal in order to remove phase dependency therefrom.

Claim 12

In a method of determining visual sensitivity as a function of image content, the step of adjusting visual sensitivity as a function of radial frequency.

Claim 13

In a method of determining threshold elevation as a function of image content of first and second images, the step of setting threshold elevation to a minimum of the threshold elevations of the first and second images.