Research BEFORE mid-2004: Research Overview

(to First Page of site)
(to Research AFTER mid-2004)
(see CV for list of publications without abstracts or pdfs)
(to Wolfson's Home Page)

LINKS TO PAGES IN RESEARCH SUMMARY:
Research Overview - YOU ARE HERE
1.0 Multiple Analyzers
1.5 Decision Stage
2.0 Complex Channels
2.5 Constant-Difference Exps.
3.0 Normalization
3.5 Early-local Nonlinearities (texture segregation)
4.0 Light adaptation


PUBLICATIONS
Chronological List with abstracts and links to pdfs for research BEFORE mid-2004
Organized List 1989 to mid-2004 (organized under the headings of Research Summary)

Four Visual Processes

Short summary: We use mathematical models to interpret data from psychophysical studies (some done by us) and neurophysiological studies (usually done by others) of visual perception. We are particularly interested in four visual processes (blue names in drawing), and also in the problem of adequately characterizing the "decision stage" (also in blue). Below the drawing you can find brief descriptions of each process, and further information is given on subsequent pages. We continue to be interested in all these processes, but in the near future we will concentrate on studies of the dynamics of the normalization process.

 

1. MULTIPLE VISUAL PATTERN ANALYZERS

Also known as "mechanisms", "channels", "filters", "pathways", etc. They analyze visual scenes into "parts" on many dimensions. One example is spatial-frequency (Fourier) channels, sometimes referred to as multiscale or multiresolution representations, where different channels are sensitive to different ranges of spatial frequency and orientation. (A pattern of narrow stripes or fine-grained texture contains predominantly high spatial frequencies. A pattern of large-sized elements contains predominantly low spatial frequencies. )

The physiological substrate probably cortical area V1

For more about Multiple Visual Pattern analyzers

1.5. The decision stage

Assumptions are necessary to get from the outputs of the processes of interest to the responses of the observer. We will refer to these assumptions as describing "the decision stage."

Multidimensional signal detection theory deals with questions of how outputs from multiple analyzers (channels, pathways, mechanisms ....) enter into this decision.

Several places in the research summaries here discuss the decision stage; links to these places are given on a page about the decision stage.

2. COMPLEX(non-Fourier, second-order) CHANNELS

Complex channels consist of small receptive fields at first stage with large fields at the second stage separated by a nonlinearity of the rectification type. Responds to low-spatial-frequency patterns of high-spatial-frequency information. "Complex channels" are often called by other names, e.g. "second-order units", or "non-Fourier mechanisms.

Physiological substrate seems likely to be cortical area V1 or V2 in our texture-segregation task, but we are not sure, and similar processes probably occur in other cortical areas

For more about Complex (non-Fourier, second-order) Channels

2.5 Constant-Difference-Series Experiments.

Constant-difference-series experiments have proved very useful in investigating both the spatial nonlinearity (see Complex channels above) and the intensive nonlinearity (now thought to be normalization rather than an early-local nonlinearity, see 3.5 and 4 below).

For description of these experiments and some typical results

3. NORMALIZATION (contrast gain control, inhibition)

In a normalization network the response of one channel/neuron is inhibited by the responses from a pool of channels and thus the single channel's repsonse is"normalized" relative to the total response from a group of channels/neurons. Normalization seems to be the intensive nonlinearity important in visual texture segregation and similar perceptual tasks.

The physiological substrate for normalization in our texture segregation tasks is probably inhibition among neurons in cortical area V1 or V2 , but normalization probably occurs in other cortical areas as well.

See more about Normalization (contrast-gain control)

3.5 Early local nonlinearities (e.g. light adaptation) in texture segregation models

A model containing an early (that is, before the channels), local nonlinearity (but no normalization) must be rejected as a complete explanation of the phenomena showing an intensive nonlinearity in texture segregation. But nonetheless the early-local nonlinearities derived by fitting this model to data can be useful summaries of certain aspects of the texture-segregation data.

Of course, nonlinear processes that occur before the channels and are relatively local do exist (e.g. light adaptation) and are of interest in themselves (see 4. below)

See more about Early-Local Nonlinearities in Texture Segregation

4. LIGHT ADAPTATION (luminance gain control)

Light adaptation allows visual system to operate throughout the vast range of luminance that occurs naturally.

The physiological substrate for light adaptation is probably largely retinal.

See more about Light Adaptation (luminance-gain control)

 

(back to top of this page)