Determining what features are critical for representing a basic shape is an important step towards understanding how the visual system perceives objects. The aim of this thesis is to determine how curvature variation around a closed-contour affects global shape processing. Specifically, the effect of points of maximum curvature (corners) on shape processing using Radial Frequency (RF) patterns is examined. RF patterns are closed-contour shapes where the radius is sinusoidally modulated. The shape’s appearance can be modulated by varying the amplitude and periodicity of the sine function. RF patterns have the advantage of being mathematically precise shapes – which can be easily manipulated to form a large number of different shapes. In this thesis it was initially intended that global shape would be assessed by measuring the size and direction of the shape after-effect following adaptation to an RF pattern. However, the first section of this thesis presents evidence which casts doubt on the suitability of shape adaptation as a tool for investigating global processes. In this section, a model is introduced which shows that the shape after-effect can be accounted for by local orientation adaptation mechanisms alone. A further limitation is the perseverance of after-effects following brief periods of adaptation (approximately 15 seconds of adaptation in 30 millisecond epochs can cause a perceptual distortion that decays slowly over 24 hours). Further research should be conducted to clarify the reliability of experiments using adaptation to investigate global shape properties. An alternative stimulus is therefore suggested for use in the investigation of global shape properties. In the remainder of the thesis, Gabor field arrays (where either a subset or all of the Gabor patches are aligned to represent a shape) are employed.
|Qualification||Doctor of Philosophy|
|Publication status||Unpublished - 2012|