ABSTRACT
Images provide important information in scientific, technical, and forensic situations, in addition to their role in everyday life. Extracting information from images acquired by digital cameras involves image processing to correct colors, reduce noise, and correct for nonuniform illumination or nonplanar views. Enhancement of image details is generally accomplished by reducing the contrast of other information in the image, so that,for example, lines and edges that make measurements of structure are more accessible. The processing steps use a variety of computer algorithms and may be performed on the pixel array, or in a different space, for example, by using a Fourier transform. Some applications, especially forensic ones, require simple comparisons, but for object identification, classification, or correlations, quantitative measurements of color or density, position, size, and shape are needed. Several possible measurement quantities are available for each category, particularly shape, for which a variety of dimensionless ratios, Fourier or wavelet coefficients, and invariant moments may be used. Interpretation of the measurements depends on the nature of the image and of the specimen or scene, for instance, whether it consists of discrete objects on a surface, a section through a complex structure, or a projection through a three-dimensional space.
