Computer interpretation of human stick figures
Martin Herman · National Conference on Artificial Intelligence · 1980
A computer program which generates context-sensitive descriptions of human stick figures is described. Three categories of knowledge important for the task are discussed: (1) the 3-D description of the figures, (2) the conceptual description of the scene, and (3) heuristic rules used to generate the above two descriptions. The program's representation for these descriptions is also discussed.