Chapter Summary

• Pattern recognition. Template theorists argue that stimuli are matched against miniature copies or templates of previously presented patterns. Unless the implausible assumption is made that there is an almost infinite number of templates to handle all possibilities, template theories seem inadequate to account for the versatility of perceptual processing. Feature theorists emphasise that any stimulus consists of specific features, and that feature analysis plays a crucial role in pattern recognition. The effects of context and of expectations are de-emphasised in most feature theories, as are the inter-relationships among features. A more adequate way of accounting for pattern recognition is provided by theories based on structural descriptions, which specify the structural arrangement of the constituent parts of a pattern.

• Object recognition. According to Marr, three main kinds of representation are involved in object recognition. The primal sketch makes use of information about light-intensity changes to identify the outline shapes of visual objects. This is followed by the 2y-D sketch, which incorporates a description of the depth and orientation of visible surfaces. It is observer-centred or viewpoint-dependent, whereas the subsequent 3-D model representation is viewpoint-invariant and provides a three-dimensional description of objects. Biederman developed this approach, assuming that objects consist of basic shapes known as geons. An object's geons are determined by edge extraction processes focusing on invariant properties of edges (e.g., curvature), and the resulting geonal description is viewpoint-invariant. Edge information is often insufficient to permit object recognition, and surface information (e.g., colour) is often more involved in object recognition than predicted by Biederman. The theories of Marr and of Biederman were designed to account for easy categorical discriminations, and viewpoint-invariant processes are less important for hard within-category discriminations.

• Cognitive neuropsychology approach. Visual agnosia can be sub-divided into apperceptive agnosia and associative agnosia. Some agnosic patients have problems in integrating information from the parts of objects. Many agnosic patients have greater problems in identifying pictures of living than of non-living objects, perhaps became pictures of living objects are more similar to each other. However, a few patients show the opposite pattern, suggesting that some of the semantic knowledge about living and non-living objects is stored in different regions of the brain.

• Cognitive science approach. Cognitive scientists have proposed connectionist models, which they have than "lesioned" to mimic the effects of brain damage on perception. The model of Farah and McClelland (1991), learned object recognition effectively, and it mimicked the double dissociation between object recognition for living and non-living objects found in patients with associative agnosia. However, the strong interconnectedness of visual and functional information in the model does not allow it to account for certain forms of visual disorder. Humphreys et al., (1995) put forward an interactive activation and competition model that accounts reasonably well for normal object recognition and for various perceptual disorders. Kosslyn et al. (1990) put forward a computer model of higher-level perceptual processes consisting of several subsystems. Lesions to various parts of the model mimicked perceptual disorders such as visual agnosia, prosopagnosia, and simultanagnosia. This model does not identify the detailed processes involved in perception.

• Face recognition. Several kinds of information can be extracted from faces, with important differences existing between familiar and unfamiliar faces. It is very rare for anyone to put a name to a face without knowing anything else about the person. There is good evidence for configural processing of faces, but there is also component processing (especially of inverted faces). Prosopagnosic patients cannot recognise familiar faces, but generally possess some implicit knowledge about them. The available evidence suggests that the difficulties of prosopagnosic patients occur because of damage to specific face-processing mechanisms rather than a general inability to make precise discriminations. Farah's two-process model distinguishes between holistic analysis and analysis by parts. Face recognition involves mainly holistic analysis, whereas reading involves mainly analysis by parts, and object recognition involves both processes.

Stop Anxiety Attacks

Stop Anxiety Attacks

Here's How You Could End Anxiety and Panic Attacks For Good Prevent Anxiety in Your Golden Years Without Harmful Prescription Drugs. If You Give Me 15 minutes, I Will Show You a Breakthrough That Will Change The Way You Think About Anxiety and Panic Attacks Forever! If you are still suffering because your doctor can't help you, here's some great news...!

Get My Free Ebook


Post a comment