In recent years, advertisers and magazine editors have been widely criticized for taking digital retouching to an extreme. Impossibly thin, tall, and wrinkle- and blemish-free models are routinely splashed onto billboards, advertisements, and magazine covers. The ubiquity of these unrealistic and highly idealized images has been linked to eating disorders and body image dissatisfaction in men, women, and children. In response, the American Medical Association recommended that publishers refrain from creating such unrealistic images, and several countries have considered legislating the labeling of retouched photos.
In response to these concerns I and Eric Kee (a graduate student in my Dartmouth lab) developed a perceptually-based metric that rates a photo on the amount by which it was altered. It is our hope that this metric can be used to inform readers by how much a photo has strayed from reality and may also encourage publishers to reduce excessive photo retouching.
Photo-editing software allows photo editors to easily alter the appearance of a person. These alterations may affect the geometry of the subject and may include slimming of legs, hips, and arms, elongating the neck, improving posture, enlarging the eyes, or making faces more symmetric. Other photometric alterations affect skin tone and texture. These changes may include smoothing, sharpening, or other operations that remove or reduce wrinkles, cellulite, blemishes, freckles, and dark circles under the eyes. A combination of geometric and photometric manipulations allows photo retouchers to subtly or dramatically alter a person’s appearance.
We have developed a metric that quantifies the perceptual impact of geometric and photometric modifications by modeling and then estimating common photo retouching techniques. Geometric changes are modeled with a dense locally linear, but globally smooth, motion field. Photometric changes are modeled with a locally-linear filter and a generic measure of local image similarity. These model parameters are automatically estimated from the original and retouched photos. Shown in the figure below, from left to right, are an original photo, the retouched photo, and a visualization of the measured geometric and photometric modifications: the measured geometric distortions are color coded from red to blue denoting large to small amounts of shape changes, the first photometric distortion embodies blurring (red) and sharpening (blue) and the second photometric distortion embodies overall changes to the face (red denotes minimal changes while blue denotes larger changes). The extent of photo manipulation is quantified with eight summary statistics extracted from these models.
In order to combine these measurements into a perceptually meaningful metric, we asked 390 people to rank 468 original and retouched photos on the degree to which they thought the person’s appearance was altered. We then used machine learning techniques to correlate these responses to our measurements and showed that we could predict perceptual judgements with a fairly high degree of accuracy. The result is a perceptually meaningful measure of photo retouching that can automatically be estimated.
Shown below, for example, are five original (top) and retouched (bottom) photos which we ranked (from left to right) with a score of 1.3, 2.4, 3.0, 3.4, 4.5. A score of 1 means minimal photo retouching and a score of 5 means significant photo retouching.
It is our hope that providing such a rating alongside published photos will better inform the public and will provide an incentive to publishers to reduce what many publich health advocates agree is an excessive amount of photo retouching.
Download: A Perceptual metric for photo retouching
- top: photo: Laflor Photography, illustration: Eric Kee
- middle: http://www.antesydespues .com.ar/en/20-famosas-sin-photoshop/
- bottom (1): http://www.flickr.com/photos/reginapagles/5559285896/in/photostream
- bottom (2): http://th05.deviantart.net/fs48/300W/i/2009/198/4/9/Beauty_retouch_9_by_hidden_silly.jpg
- bottom (3): http://www.befter.net/user/ThalesRC/beft/kim-kardashian-before-and-after-photoshop/
- bottom (4): http://www.antesydespues.com.ar/20-famosas-sin-photoshop/
- bottom (5): http://www.flickr.com/photos/pauloarrivabene/2503732663/in/set-72157605134336657