Abstract: Various embodiments contemplate systems, architectures and methods for extracting and selecting headshots of human or non-human entities from catalogs of images of such subjects. The methods described may find and extract faces from within groups of subjects, verify that the extracted faces correspond to the desired subject, determine cropping or masking regions, or both, of rectangular, circular, elliptical or some other geometry to provide an easily recognized image of the desired subject, expand the output image by synthesizing pixels as may be needed for a desired cropping or masking region, select preferred images among a collection of images of the desired subject, and perform other useful functions. The resulting output images may be in either direct form, reference form or both forms.

Abstract: Various embodiments contemplate systems, architectures and methods for extracting and selecting headshots of human or non-human entities from catalogs of images of such subjects. The methods described may find and extract faces from within groups of subjects, verify that the extracted faces correspond to the desired subject, determine cropping or masking regions, or both, of rectangular, circular, elliptical or some other geometry to provide an easily recognized image of the desired subject, expand the output image by synthesizing pixels as may be needed for a desired cropping or masking region, select preferred images among a collection of images of the desired subject, and perform other useful functions. The resulting output images may be in either direct form, reference form or both forms.