A method for retrieving from a digital photograph of a human face of the information sufficient for its recognition
Abstract
A method for retrieving from a digital photograph of a human face of the information sufficient for its recognition
Incoming article date: 19.08.2014In this report, we propose an approach to the automatic extraction of such components of the image, that are most useful for human observers to solve the recognition task. This approach is based on the concepts of psychophysiological mechanisms of grouping the simple features at the preattentive stages of visual processing. A method is based on digital image processing by a software implemented model "filter-rectify-filter". The extraction of the most informative areas of the photograph is based on finding local maxima of the model "outputs". The calculation results indicate that the areas of the image containing spatial modulations of the local features may be very useful for the recognition of faces.
Keywords: Second order visual mechanisms, image filtering, regions of interest, face identification, computer simulation