elkins@topaz.RUTGERS.EDU (George Elkins) (12/12/86)
I am interested in David Marr's work on artificial vision and would like pointers to articles and books further discussing this work and its effects in the field of artificial vision. Specifically, I would like to find out what psychophysical evidence supports or contradicts Marr's model of early visual processing (or any other model of vision that is similar or utilizes Marr's analysis of a grey-level intensity array to extract zero-crossings and other cues for the construction of a primal sketch, 2 1/2 D sketch, etc.) What are the differences between the human visual system's processes and Marr's work? Isn't the entire concept of using any gray-level INTENSITY matrix contradictory to what is known of the human visual system, which seems from the beginning to only detect CHANGES in intensity? For example, a stabilized retinal image "disappears." (I've heard the term "Photometer Metaphor" used in this sort of criticism of the usage of a grey-level intensity array, but I can't exactly remember where I read or heard this.) I assume that Marr intended to partially make up for this in the application of the "Laplacian of the Gaussian" to get zero-crossings in the image. How closely does this correspond to early processing in the human visual system? Pardon my ignorance, I'm new to this field. George Elkins rutgers!topaz!elkins or maybe even rutgers!caip!elkins