Method for digital image class detection
View Patent ↗A system and method for analyzing and classifying digital images as color, black-and-white, black-and-white with spot color, monochromatic, or monochromatic with spot color is provided. According to an embodiment of the present invention, a digital image is received and selected pixel values thereof are transformed from RGB color component values to a lightness component value and two color difference component values. The selected pixels are analyzed with respect to neutrality based upon the two color difference values and are grouped according to lightness value. The sample population distributions of each of the two color difference values are statistically analyzed for each group of pixels, and an image classification is determined by examination of the resulting statistics and derivatives thereof.
1. A method implemented at least in part by a computer system, the method for determining a color classification of a digital image comprising pixels each having at least three color values, and the method comprising the steps of:
converting the color values of at least some of the pixels to a lightness value, a first color difference value, and a second color difference value, the at least some of the pixels being selected pixels;
assigning each of the selected pixels to one of a plurality of lightness level groups based at least upon the selected pixel's lightness value;
using the computer system to calculate, for at least some of the lightness level groups (“selected LLGs”), a first mean of first color difference values and a second mean of second color difference values of corresponding selected pixels;
using the computer system to calculate, for each of the selected LLGs, a mean color hue angle value from the first mean of the first color difference value and the second mean of the second color difference value; and
determining whether the digital image is monochromatic based at least upon (a) an analysis of a distribution of the first and second color difference values in each of the selected LLGs with respect to the calculated means, or derivatives thereof, and (b) an analysis of the distribution of mean color hue angle values for all the selected LLGs, or derivatives thereof.
2. The method of claim 1 , wherein the analysis of the distributions of the first and second color difference values includes a calculation of a range, a minimum value, a maximum value, a variance, or a standard deviation of each distribution.
3. The method of claim 1 , wherein the analysis of the distributions of the first and second color difference values includes a calculation of a range, a minimum value, a maximum value, a variance, and a standard deviation of each distribution.
4. The method of claim 1 , wherein the digital image is first analyzed to determine if all of the selected pixels are monochrome pixels of a white, black, or a shade of gray between white and black, based at least upon an analysis of the first and second color difference values for all selected pixels, and wherein the determining step determines whether the digital image is grayscale monochromatic.
5. The method of claim 1 , wherein the digital image is first analyzed to determine whether each of a majority of the selected pixels is a monochrome pixel of a white, black, or a shade of gray between white and black, based at least upon an analysis of the first and second color difference values for each selected pixel, and wherein the determining step determines whether the digital image is grayscale monochromatic with a spot accent color.
6. The method of claim 1 , wherein the selected pixels comprise only a central portion of the digital image.
7. The method of claim 1 , wherein the number of pixels in the digital image is reduced in size prior to any analysis performed as part of the determining step.
8. The method of claim 1 , wherein results of the analysis of the distributions of the first and second color difference values in each of the selected LLGs are weighted as a function of an average lightness level for each of the selected LLGs.
9. The method of claim 1 , wherein the analysis of the distributions of the first and second color difference values excludes distributions from under-populated selected LLGs.
10. The method of claim 1 , wherein a first statistical measure of normality is calculated from the distribution of the first color difference values in each of the selected LLGs, wherein a second statistical measure of normality is calculated from the distribution of the second color difference values in each of the selected LLGs, and wherein the determining step determines whether the digital image is monochromatic with spot accent color based at least upon (a) an analysis of the first statistical measure of normality and the second statistical measure of normality values in each of the selected LLGs, or derivatives thereof, and (b) an analysis of the distribution of mean color hue values for all the selected LLGs, or derivatives thereof.
11. A computer-accessible memory system storing instructions configured to cause a data processing system to implement a method for determining a color classification of a digital image comprising pixels each having at least three color values, wherein the instructions stored on a computer-readable medium comprise:
instructions for converting the color values of at least some of the pixels to a lightness value, a first color difference value, and a second color difference value, the at least some of the pixels being selected pixels;
instructions for assigning each of the selected pixels to one of a plurality of lightness level groups based at least upon the selected pixel's lightness value;
instructions for calculating, for at least some of the lightness level groups (“selected LLGs”), a first mean of first color difference values and a second mean of second color difference values of corresponding selected pixels;
instructions for calculating, for each of the selected LLGs, a mean color hue angle value from the first mean of the first color difference value and the second mean of the second color difference value; and
instructions for determining whether the digital image is monochromatic based at least upon (a) an analysis of a distribution of the first and second color difference values in each of the selected LLGs with respect to the calculated means, or derivatives thereof, and (b) an analysis of the distribution of mean color hue angle values for all the selected LLGs, or derivatives thereof.
12. A system comprising:
a data processing system; and
a memory system communicatively connected to the data processing system storing instructions configured to cause the data processing system to implement a method for determining a color classification of a digital image comprising pixels each having at least three color values, wherein the instructions stored on a computer-readable medium comprise:
instructions for converting the color values of at least some of the pixels to a lightness value, a first color difference value, and a second color difference value, the at least some of the pixels being selected pixels;
instructions for assigning each of the selected pixels to one of a plurality of lightness level groups based at least upon the selected pixel's lightness value;
instructions for calculating, for at least some of the lightness level groups (“selected LLGs”), a first mean of first color difference values and a second mean of second color difference values of corresponding selected pixels;
instructions for calculating, for each of the selected LLGs, a mean color hue angle value from the first mean of the first color difference value and the second mean of the second color difference value; and
instructions for determining whether the digital image is monochromatic based at least upon (a) an analysis of a distribution of the first and second color difference values in each of the selected LLGs with respect to the calculated means, or derivatives thereof, and (b) an analysis of the distribution of mean color hue angle values for all the selected LLGs, or derivatives thereof.