The Image Quality Circle (IQC) is a process for managing the image quality of imaging products. It consists of four elements illustrated by the rounded rectangles in the illustration:
Connecting the four elements are three connecting links, represented by the rectangles in the illustration:
The double arrows indicate that the process works in both directions. You can go around the Image Quality Circle in both clockwise and counter clockwise directions.
The large red arrow connecting the Customer (Image) Quality Rating with the Technology Variables represents the historical process. This consists of generating sample images, by varying the Technology Variables, and getting Customer Image Quality Ratings by a suitable psychometric scaling method. The results are limited. You only know the quality rating for one set of the variables. It is not often generalizable.
Connecting the Customer Image Quality Rating with the Technology variables, via the IQC, provides an alternative and more comprehensive route toward Image Quality Management.
Technology Variables - The things we manipulate. These are the almost endless list of parameters or variables, hardware, firmware and software, that an imaging technologist trades-off to produce a known level of image quality.
Physical Image Parameters - The things we measure. The quantitative functions and parameters that we ascribe to images. Generally, measured quantities that physically characterize an image or imaging system (image physics). Examples are: spectral reflectance factor, optical density, imaging element position, Wiener spectra, etc.
Customer Perceptions - The things people see. The components, or perceived attributes of images. Often the names of these perceptions have the suffix "ness", such as darkness, sharpness, colorfulness, etc. A shorthand notation that is used in the Image Quality Circle is "nesses."
Customer Image Quality Rating - The quality number. This number is determined by having customers, or customer surrogates, scale sample images and express an image quality judgment. An example is a "50" on a 0 to 100 scale of image quality.
System/Image Models - The formulas, algorithms, computer code, and neural nets, etc., that predict Physical Image Parameters from Technology Variables. The System Model inputs are values of technology variables, and the output(s) is a value that directly relates to a physical measurement of an image. An example of a System Model might be the inputs of dot spacing and diameter of ideal circular ink dots that yields the RMS variation of the boundary of an object.
Visual Algorithms - Computer algorithms, neural nets, etc. that link the Physical Image Parameters to the Customer perceptions, the "nesses." Functionally these algorithms map some physical measurement of an image to an attribute perceived by the customer, a "ness". For example: the physical measurement of the RMS variation of an object boundary, the Physical Image Parameter, in the above example; and a number representing the percept of the boundary raggedness or roughness.
Image Quality Models - Empirical, often statistical, equations that link the Customer Percepts, the "nesses", with the numbers representing image quality.
For further information email Peter Engeldrum
Back
Copyright 2009 Imcotek