Visual Information-Processing in the Perception of Features and Objects
Abstract
This research supported by my grant from AFOSR this year completed some of the projects outlined in the first annual report and initiated some new ones. The focus remained on the visual processing of features and objects, the role of spatial attention and the representation of complex visual patterns in perception and memory. Studies of visual search explored the coding of features at potentially more abstract levels than simple luminance filters: Among the features studied were orientation (for lines, dot pairs and edges), orientation and size (for shapes whose boundaries were defined by discontinuities of texture, motion and depth), and illusory contours. Another series of experiments tested the mechanisms underlying the coding of feature conjunctions, using evidence from search latencies and illusory conjunction errors. The results led to a proposed revision of my earlier feature integration theory. Two other studies looked at memory for visual patterns. One studied the effects of prolonged practice (thousands of trials) on the coding of visual patterns. At the other extreme, an experiment explored the effects on memory just one to five presentations of similar patterns. In both cases, we found striking specificity in the coding of these meaningless shapes and in their effects on subjects' later experiences with the same stimuli. Patterned conjunctions of lines, once formed, appear to persist in visual memory and to facilitate their later re- perception provided that the same task and context are also reinstated. Keywords: Attention.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 05, 1989
- Accession Number
- ADA206948
Entities
People
- Anne Treisman
Organizations
- University of California, Berkeley