Use this URL to cite or link to this record in EThOS:
Title: Psychophysical investigations of visual density discrimination
Author: Macpherson, Timothy
Awarding Body: University of London
Current Institution: University College London (University of London)
Date of Award: 1991
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Work in spatial vision is reviewed and a new effect of spatial averaging is reported. This shows that dot separation discriminations are improved if the cue is represented in the intervals within a collection of dots arranged in a lattice, compared to simple 2 dot separation discriminations. This phenomenon may be related to integrative processes that mediate texture density estimation. Four models for density discrimination are described. One involves measurements of spatial filter outputs. Computer simulations show that in principle, density cues can be encoded by a system of four DOG filters with peak sensitivities spanning a range of 3 octaves. Alternative models involve operations performed over representations in which spatial features are made explicit. One of these involves estimations of numerosity or coverage of the texture elements. Another involves averaging of the interval values between adjacent elements. A neural model for measuring the relevant intervals is described. It is argued that in principle the input to a density processor does not require the full sequence of operations in the MIRAGE transformation (eg. Watt and Morgan 1985). In particular, the regions of activity in the second derivative do not need to be interpreted in terms of edges, bars and blobs in order for density estimation to commence. This also implies that explicit coding of texture elements may be unnecessary. Data for density discrimination in regular and random dot patterns are reported. These do not support the coverage and counting models and observed performance shows significant departures from predictions based on an analysis of the statistics of the interval distribution in the stimuli. But this result can be understood in relation to other factors in the interval averaging process, and there is empirical support for the hypothesized method for measuring the intervals. Other experiments show that density is scaled according to stimulus size and possibly perceived depth. It is also shown that information from density analysis can be combined with size estimations to produce highly accurate discriminations of image expansion or object depth changes.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available