Measuring image quality automatically requires that feature locations be known and predictable. This predictability is necessary for successful application of both large field of view, low magnification image analysis (such as mottle) and small field of view, high magnification image analysis (such as line and dot quality) in a machine-vision based system. Variations in sample placement and in the registration of images on the samples produce systematic errors in absolute feature locations that require compensation.Compensation for systematic image placement errors is straightforward. Macroscopic image or sample placement variations can be corrected automatically by using a large field of view camera and an image offset calculation. However, even after the macroscopic image placement correction has taken place, microscopic position variations can still exist. These variations can wreak havoc with high magnification measurements that rely on specific region of interest placement on specific features within the field of view. In some cases, dynamic location can be used for small adjustments within the field of view of the high magnification camera. This combination of macro and micro adjustment works well for most samples where absolute or relative feature positions are known.Difficulties arise when positional variations of features are not systematic or rectilinear. Certain types of samples are susceptible to local or global deformation, which can change the relative position of features. For example, cockled paper, or ink jet samples with highly saturated areas are prone to local deformations, while textiles exhibit major topological variations as well as global deformations that result in non-rectilinear distortions caused by the fabric weave. Although topological variations in samples printed on paper and other paper-like substrates can be largely compensated for through the use of a strong vacuum to hold a sample flat, there are residual, non-systematic positional changes of features that can interfere with traditional automated measurement techniques.This paper will present a series of specific test target design considerations and testing methods that can be used to enable automated image quality measurement of distorted image samples.