Image Quality Concepts

Abstract

We present several concepts useful in the specification, evaluation, and control of image quality for digital capture. We start by introducing the terminology and application of image quality models. The value of these models, however, lies in the ability to connect them to characteristics of images or imaging systems. Translating image quality into technology selection and design can be done when we have a physical understanding of design and performance. We describe methods for the objective measurement of several key imaging characteristics. They are then related to current practice by reference to several established imaging performance standards. These system‐based methods are often complemented by applying human vision models. Vision models are also part a larger set of techniques collectively known as Image Quality Assessment (IQA). IQA is based on extracting measures directly from digital images, particularly useful for processing paths which adapt to image content.

Camera System Performance Derived from Natural Scenes

“…The Modulation Transfer Function (MTF) [1][2][3] is the primary imaging performance measure used for sharpness and resolution evaluation, standardized in the ISO12233 [4]. It is equal to the modulus of the Fourier transform of the Point Spread Function (PSF), the latter being defined, for a continuous system, as the image of a point source.…”

“…There are various methods for MTF evaluation, including imaging of point sources, slits (lines), 'perfect' edges, series of sinusoids, and stationary stochastic noise fields, or patterns of known spectral contents [3,5,6]. Each of these comes with several implementations, different advantages and disadvantages, as well as associated measurement errors.…”

Texture MTF from images of natural scenes

“…Often this might not be possible, as in the majority of camera phones. Also, it might be of greater interest to evaluate the performance of the device in its normal use [15]. The exponential growth of the camera phones' industry has led to the development of MTF measurements that rely on test images with content that better resembles natural textures, with power spectra similar to these of real scenes [16].…”