In addition to the quality of an original (transparency or reflection) preferred halftone reproduction is also dependent on scene classification. That is, how the tones of the original are rendered. If the majority of the picture area contains light tones, it is referred to as a high key image. Should the majority of picture area be confined to dark tones, it is referred to as a low key image. When the picture area is composed of the entire tone scale (white, grays, and black), it is referred to as a normal key image. The problem of how to objectively classify images into high, normal, and low key scene classification still continues in the graphic arts. The intent of this thesis was to deter mine if a scanning device could be attached to a HCM 286 Color Scanner and applied to a correlation study between the classification and ranking of images by observers and that of an objective measure of the images. One experiment was devoted to psycho-physical testing in which several black and white photographs were subjectively classified and ranked into high, normal, and low key image types. Installation of the scanning device revealed complications that made it necessary to go ahead and obtain the objective data by using manual density measurements and a computer to derive the images' tone distribution curves (TDCs). The TDC is a statistical representation of the images' tone information; not a test object. Because of the tediousness of manual density measurement brought about by the inavailability of the scanning device, only eight images were measured objectively. Based on the complexity of their TDC shapes, the existence of a real relationship between subjective and objective scene classification cannot be inferred presently. It is believed that this relationship can be found when a larger sample size of objective data is available. However, it is only reasonable to consider a larger sample size when a less tedious method becomes available, such as the scanning device. Discussions and recommendations are also given for further investigation between sequential viewing (one image at a time) and simultaneous viewing (all images at one time).
Library of Congress Subject Headings
Optical scanners; Images, Photographic; Photomechanical processes
Department, Program, or Center
School of Photographic Arts and Sciences (CIAS)
Campbell, Craig, "Correlation Study Between Subjective and Objective Scene Classifications" (1977). Thesis. Rochester Institute of Technology. Accessed from
RIT – Main Campus