Systems, methods, and articles of manufacture for generating an equalized image using signature standardization from Weibull space

摘要

Systems, articles of manufacture, and methods for generating an equalized image using properties of the two-dimensional Weibull distribution.

主权项

1. A method for image equalization using an electronic processor, comprising:
electronically determining a mean intensity value of a plurality of pixels of an input digital image having a plurality of possible pixel intensity values, wherein each of said plurality of pixels has an initial pixel intensity value; electronically generating a first line fit of a first amplitude region of a cumulative distribution function of said initial pixel intensity values of said plurality of pixels mapped in a Weibull space; electronically generating a second line fit of a second amplitude region of said cumulative distribution function of said initial pixel intensity values of said plurality of pixels mapped in said Weibull space; electronically determining an intersection point of said first line fit and said second line fit in said Weibull space; electronically determining an abscissa value of said Weibull space corresponding to said intersection point; electronically generating intermediate pixel intensity values for said plurality of pixels by:
electronically generating intermediate intensity values of the pixels of said plurality of pixels that have an intensity value lower than said mean intensity value rounded down to a nearest rounded mean intensity value of said plurality of possible pixel intensity values by assigning each of said plurality of pixels having an intensity value lower than said nearest rounded mean intensity value a value equal to the value of said plurality of possible pixel intensity values that sequentially immediately precedes said nearest rounded mean intensity value when said plurality of possible pixel intensity values are sorted from least to greatest;electronically generating intermediate intensity values of the pixels of said plurality of pixels having an intensity value greater than a value of the exponential of said abscissa value rounded up to a nearest rounded exponential of said abscissa value of said plurality of possible pixel intensity values by assigning each of said plurality of pixels having an intensity value greater than said nearest rounded exponential of said abscissa value a value equal to the value of the possible intensity values that sequentially immediately follows said nearest rounded mean intensity value when said plurality of possible pixel intensity values are sorted from least to greatest; andelectronically generating intermediate intensity values of all remaining pixels by assigning each of all other pixels of said plurality of pixels an intensity value equal to its said initial pixel intensity value; electronically determining an equalized intensity value for each of said plurality of pixels according to an equalization transfer function mathematically described in the following equation, where ν represents said intermediate intensity values, h(ν) is the equalized intensity value corresponding to intermediate intensity value ν, CDFintermediate(ν) is a cumulative distribution function value of intermediate intensity value ν, CDFintermediate—min is the smallest CDF value of the cumulative distribution function of said intermediate intensity values, L is a number of gray levels in said input digital image, M is a number of rows in said input digital image, and N is a number of columns in said input digital image:h⁡(v)=round(⁢CDFintermediate⁡(v)⨯(M⨯N)-CDFintermediate_min⨯(M⨯N)(M⨯N)-CDFintermediate_min⨯(M⨯N)⨯(L-1)); and electronically generating an equalized image by assigning each of said plurality of pixels a final intensity value equal to its corresponding said equalized intensity value.