Pattern recognition studies in the biomedical sciences

Abstract

The biomedical sciences characteristically deal with huge masses of data, which must be organized, reduced, analyzed, and generally processed in many different ways. Much of this data is in the form of pictures: photomicrographs, electron micrographs, X-ray films, Schlieren photographs, X-ray diffraction patterns, autoradiographs, time-lapse films, cineradiographs, or the like. Individual pictures hold a great wealth of precise numerical information, such as the morphological and structural characteristics of lengths, areas, volumes, and densities. From sequences of pictures, quantitative results can be derived, such as the kinematic and dynamic characteristics of trajectories. Such pictures relate to almost every field of biomedical research: chromosome karyograms in cytogenetics, angiogram cineradiographs in cardiology, Schlieren photographs in ultracentrifugal molecular-weight determinations, autoradiographs of polymorphonuclear leukocytes in the study of leukemia, Golgi-stained neuron photomicrographs in the study of the ontogeny and phylogeny of the brain, X rays of bones in studies of calcium density distribution in orthopedic diseases, X rays of epiphysial plates of the hand in investigations of accurate physiological age, X-ray crystallographic plates in protein structure determination, electron micrographs in the investigation of the fine structure of virus particles, motion pictures of marine crustaceans in the detection of their sensitivity to polarized light, tissue-culture time-lapse films in the investigation of cancer-cell motility, and many others.