Celebrity Caricature

Some, on the other hand, argue that caricature varies depending on the artist and cannot be captured in a single definition. Their system uses machine learning techniques to automatically learn and mimic the style of a particular caricature artist, given training data in the form of a number of face photographs and the corresponding caricatures by that artist. The results produced by computer graphic systems are arguably not yet of the same quality as those produced by human artists. For example, most systems are restricted to exactly frontal poses, whereas many or even most manually produced caricatures (and face portraits in general) choose an off-center "three-quarters" view. Brennan's caricature drawings were frontal-pose line drawings. More recent systems can produce caricatures in a variety of styles, including direct geometric distortion of photographs.
In a lecture titled The History and Art of Caricature (September 2007, Queen Mary 2 Lecture theatre), the British caricaturist Ted Harrison said that the caricaturist can choose to either mock or wound the subject with an effective caricature. Drawing caricatures can simply be a form of entertainment and amusement - in which case gentle mockery is in order, or the art can be employed to make a serious social or political point. A caricaturist draws on (1) the natural characteristics of the subject (the big ears, long nose or whatever); (2) the acquired characteristics (stoop, scars, facial lines etc.); and (3) the vanities (choice of hair style, spectacles, clothes, expressions and mannerisms).
Ramachandran and Hirstein suggested that caricature is related to peak shift. In the peak shift effect, animals sometimes respond more strongly to exaggerated versions of the training stimuli. For example, if a rat is trained to respond to a rectangle of a particular aspect ratio, and to avoid a square, when later presented with several rectangles it will prefer the one with the most elongated aspect ratio (this being the one that is most different from the square) rather than the original rectangle used in training. Ramachandran and Hirstein speculated that cells in a monkey brain that respond to particular faces would respond more strongly to caricatured versions of the face. This effect has been confirmed in FMRI experiments by Tsao.