On Detection Structure and Representation of Niultiscale Low-Level Image NARENDRA AHUJA Beckman Urbana InstLtute -Champaign and Department L,lslon.al. of Electrzal ulue edu) and Computer Engtneermg, Unmerslty of Illlnou at ( ahuJci[~ The objective of computer vision is interpretation of visual images. Any datainterpretation task of such magnitude requires models of the data. For example, in speech the audio signal is parsed into phonemes, which are successively merged into increasingly complex units and eventually into an interpretation, often with feedback from higher levels. Another example is hierarchical interpretation of computer programs in a given language through the use of grammars. In image data, analogues of phonemes and characters primitives that correspond compress to structural the data to a any manageable size without eliminating possible final interpretations. Because images and more complex are significantly larger than speech signals, a data reThe low- capability for initial, bottom-up duction is even more critical. level structure would serve image abstraction and help archical, tion, for enforcing involving closed-loop example, as a lossless initiate hier- image interpretafor recognition by a priori semantic constraints part whole relationships. This note is not concerned with interpretation processes, it describes some desirable characteristics of strategies for the detection
/lp/association-for-computing-machinery/on-detection-and-representation-of-multiscale-low-level-image-BmGoW7uxVJ