The recognition of incomplete contour and half-tone figures
Shelepin, Yuri and Vahromeeva, O. and Harauzov, A. and Pronin, Sergey and Krasilnikov, N. and Foreman, Nigel and Chikhman, V. N. (2004) The recognition of incomplete contour and half-tone figures. Perception, 33 . ISSN 0301-0066
Full text is not in this repository.
This item is available in the Library Catalogue
The recognition of incomplete figures of everyday objects in fovea and beyond was the aim of our project. For fifteen subjects with high visual acuity, aged 18 to 24 years, we measured the necessary minimum of randomly presented parts of the contour. Fixation point was in the centre of the figure corresponding to the centre of the screen. The figure sizes were varied from 5 min of arc up to 20 deg by using different viewing distances. The contour minimum for recognition corresponds to figure size equal to foveola size. Beyond 20 min of arc, up to the resolution limit of 5 min of arc, recognition decreased linearly. For 0.5 to 5 deg sizes the threshold was relatively invariant to the size. The psychophysics and event-related potentials were investigated within the frequency range from 1 to 45 Hz. The relative perceptual availability of global and local features in sparse patterns was investigated during different stages of learning. Interpolation algorithms were proposed and tested in additional experiments. We measured the threshold of incomplete natural half-tone scene recognition. In these experiments, incompleteness was produced with wavelet transforms. The internal noise limitation of recognition was measured. A match filter model of the perception of incomplete figures is proposed.
European Conference on Visual Perception.
22-26 Aug, 2004.
|Research Areas:||A. Middlesex University Schools and Centres > School of Health and Education > Health & Education|
|Deposited On:||04 Dec 2009 05:26|
|Last Modified:||10 Jan 2014 05:57|
Repository staff only: item control page
Full text downloads (NB count will be zero if no full text documents are attached to the record)
Downloads per month over the past year