Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.660823
Title: Motion-deblurring mechanisms of human visual perception
Author: Pääkkönen, Ari Kullervo
Awarding Body: University of Edinburgh
Current Institution: University of Edinburgh
Date of Award: 1993
Availability of Full Text:
Access through EThOS:
Full text unavailable from EThOS. Please try the link below.
Access through Institution:
Abstract:
The temporal integration period of the visual system for both stationary and moving objects is known to be over 100 ms in daylight. One might expect from this that motion blur should considerably degrade our percepts of moving objects. However, we are able to see moving objects with clarity. Some authors have suggested that there are special motion-deblurring mechanisms to prevent motion degradation. The proposed mechanisms are motion-tuned: integration follows the motion path of the object. The aim of this work was to study the properties of visual motion blur and their implications for possible deblurring mechanisms. In the main experiment, blur discrimination thresholds for moving, Gaussian-blurred edges were measured in order to separate the effects of motion blur and static spatial blur. The data was modelled by assuming a linear transform of the physical image of the edge to its neural representation. The results show that motion produces equivalent spatial blur. The velocity dependence of this blur is linear, and its extent can be predicted by a temporal impulse response with a standard deviation of about 5 ms in normal room light, which gives an estimate of about 0 to 225 ms for the blur-producing motion integration time. Supported by the results of two-dot resolution experiments, it is proposed that this blur results from camera-like summation and not from the use of larger spatial filters for moving than for stationary objects. When the author repeated the main experiment six months later after many other experiments mainly with small reference blurs, changes in thresholds and model parameters were found that indicate low-level learning and neural plasticity in the blur discrimination system. An even more surprising result came from the experiment where the contrast polarity in relation to the direction of motion was changed opposite to that in the main experiment. Learning effects for moving edges were not transferred from one polarity to the other, indicating that spatial analysis of moving objects may be served by two separate subsystems, possibly related to the on and off systems of the visual pathway.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.660823  DOI: Not available
Share: