— Archives
“Lena Söderberg,
Jennifer Knoll
& Jennifer Lopez”
by Marie Lechner [EN/FR]

Publié le


They have marked the history of tech, in the field of image processing and manipulation, and the invention of image search engines. These three images highlight how the male gaze pervades our vision machines. They also help shed light on contemporary debates on race and gender biases1 encoded in algorithms and how current developments in artificial intelligence perpetuate a certain stereotyped view of the world.

Lena, First Lady of the internet

Image : Lena_(test_image).png
Function: test image for image processing algorithms
Year: June or July 1973
Description: a 512 x 512-pixel scan made by Alexander Sawchuck
and two colleagues for the USC-SIPI image database.

Lena Söderberg, Playboy

Lena Söderberg, « playmate of the month » in Playboy, detail, november 1972, photography by Dwight Hooker. 512 x 512-pixel scan made by Alexander Sawchuck and two colleagues for the USC-SIPI image database.

The image shows the face of a young woman in a feathered hat, bare-shouldered and posing half-turned from a mirror. This 512 x 512-pixel scan known as Lena (sometimes spelled Lenna) has been one of the most used images for image processing since 1973, due to its highly specific features. Lena holds a combination of details, flat regions, textures and shadows, allowing to test a wide variety of filtering, segmentation and compression algorithms.

This is one of two reasons why the image of Lena, which laid the foundations of the JPEG and MPEG formats, has become an industry standard, according to David C. Munson, editor in chief of the scientific journal IEEE Transactions on Image Processing2, specialized in image processing. The second is that “the Lena image is a picture of an attractive woman. It is not surprising that the (mostly male) image processing research community gravitated toward an image that they found attractive.”3

The image of Lena was scanned in June or July 1973 from an image of larger dimensions. This distorted version, cropped at the shoulder, is a fragment of a full-length portrait published as the famous centerfold in the men’s magazine Playboy – that of Lena Söderberg, a young Swedish woman, Miss November 1972, posing in her birthday suit for the photographer Dwight Hooker.

The story is told in-depth by Jamie Hutchinson in the May/June 2001 professional newsletter of the IEEE Professional Communication Society.4 Alexander Sawchuk and his colleagues from the SIPI (Signal and Image Processing Institute), engineers at the University of Southern California, were seeking to replenish their stock of test images: they were looking for something “glossy” and “they wanted a human face”. Someone showed up just then with the issue of Playboy. The engineers tore away the upper third of the image so that it would fit around the drum of their Muirhead wirephoto scanner. They were seeking to produce a 512 x 512-pixel image with this scanner, which had a resolution of 100 lines per inch. They thus only digitized 13 cm of the photo, removing its most risqué elements.

Other researchers quickly tested their own algorithms by using the photo of Lena, leading to the image’s widespread diffusion over the years. Lena became an icon within the community, “an information age Madonna” in the words of Jamie Hutchinson: “Lena became for the engineers something like what Rita Hayworth was for US soldiers in the trenches of World War II,” he writes.

Lena Söderberg, who had in the meantime returned to live in her native Sweden, had no idea that she had become a laboratory rat. Upon being recruited by Playboy, the young Swede was doing some modeling in Chicago; she would later work as a Kodak model in Rochester, one of the company’s “Shirleys”, beautiful white women whose images were used to calibrate color film (named for the first woman to have held the position, Shirley Page.) “The cover of the 7700 instruction manual shows Lena’s wide-eyed face superimposed on an image of the photocopier, as if she came along with the packaging, the girl in the machine,” comments Linda Kinstler in an article for Wired.5

The fact that Lena’s image proliferated at this exact moment in history is far from a coincidence, according to the historian Marie Hicks, author of Programmed Inequality; this was the time when the women who had worked as “computers” (prior to its current signification, the term was used to refer to people who performed calculations by hand) during the first half of the 20th century were leaving the tech industry in droves. The image of the profession, long considered as a subaltern and feminine activity, gradually and deliberately transformed into a high-level, scientific and male-dominated field.

Lena became the symbol of the profession’s unwelcoming stance toward women. In 1996, David C. Munson, current president of the Rochester Institute of Technology, wrote in his “Note on Lena”6: “I have heard feminists argue that the image should be retired.” In 1997, Sunny Bains, the editor of the journal Photonics, decided to ban Lena from the publications she was editing: “it is easy to feel isolated when you are a woman working in a field dominated by men,” she wrote in Electronic Engineering Times in May 1997. “Seeing provocative images of women in manuals contributes to this feeling of non-inclusion.” Despite regular protests, the image of Lena, “quite possibly the world’s most studied image since the Mona Lisa,”7 remains ubiquitous. “The prolific use of Lena’s photo can be seen as a harbinger of behavior within the tech industry,” writes Emily Chang in her book Brotopia. The moment when the centerfold of Lena was torn and scanned marked, for her, “tech’s original sin”. Chang also appears in the documentary Losing Lena, which advocates for putting an end to the use of Lena’s image in technological research. She symbolizes a certain attitude within the field, one that perpetuates the underrepresentation of women, as well as the gender and racial biases now highlighted in the development of artificial intelligence. Lena, 69, appears in the film: “I left modeling long ago. The time has come for me to get out of tech as well.”

Jennifer in Paradise, the first photoshopped image

Image : Jennifer in paradise.tiff
Function: test image for demos of the image processing software Photoshop
Date: August 1987
Source: the photo presented here is Jennifer in Paradise.jpeg
the version reconstructed in 2013 by the artist Constant Dullaart
from the original vacation photo taken by John Knoll.

Constant Dullaart ; Jennifer in Paradise ; Photoshop ; Jennifer Knoll ; John Knoll ; Bora Bora

Jennifer_in_Paradise, restored digital image re-distributed online with steganographically encrypted message, 2013 © Constant Dullaart

A woman seen from behind sitting topless on a beach, gazing into the distance. White sand, turquoise water, the lush island of To’opua on the horizon. We cannot see her face, but we know her name: Jennifer. The shot was taken in 1987 by her boyfriend during a romantic getaway in Bora Bora. In The Guardian, Jennifer says, “it was a truly magical time for us, my husband actually proposed to me later on in the day.” This is no doubt why John Knoll, who with his brother co-created Photoshop, the famous retouching software that would irrevocably alter our vision of the world, named this image “Jennifer in Paradise”. The highly personal photo became the first image to be publicly transformed by the most influential image manipulation software. This vacation image was distributed with early editions of what would become Photoshop, and a private moment on the beach became fodder for playing around for a number of people, who could clone Jennifer, make her disappear from the image, apply filters to deform her, and more.

In contrast with the image of Lena, she was difficult to find online. The Dutch artist Constant Dullaart discovered Jennifer’s nude back in a documentary devoted to the software’s creators. He was surprised that this “historic” photo, the first image to be publicly altered by the most influential image manipulation program, was nowhere to be found on the web, as he deplored in his 2013 “Letter to Jennifer Knoll”, in which he asks Jennifer if she would make this photo available. In a way, Jennifer was the last woman to inhabit a world in which the camera doesn’t lie. While Dullaart never received any reply or the rights, he took the time to reconstruct the image using screen shots from the video, and placed it back into circulation online in the form of a JPEG file.

Contacted by The Guardian, John Knoll explained that at the time, there were few digital images available to show off the capabilities of their new software. During a visit to Apple’s Advanced Technology Group Lab, Knoll had the opportunity to use one of the scanners, also rare at the time. The only image he had with him was this 10 x 15 cm print of his wife in Tahiti. Taken with an analogue camera, then scanned by a Sharp JX-450, several floppy discs were needed to stock this “high resolution” photo. Jennifer in Paradise thus became the first color image used to demonstrate the software’s abilities and to solicit new clients.

“It was a good image to do demos with,” Knoll recalls in The Guardian, which published the original photo to illustrate the article. “It was pleasing to look at and there were a whole bunch of things you could do with that image technically.”

“There is something absurd about using an image of his all but nude future wife as an object to manipulate and duplicate. This is exactly how Photoshop is used today by fashion magazines: women’s bodies are objectified and retouched to mask imperfections,” Constant Dullaart analyzes. “This tool illustrates the extent to which our software and tech environment is the fruit of a capitalist society dominated by men.”

3. Jennifer Lopez and her green dress behind Google Images

Images: Jennifer Lopez green dress 2000
Date: February 23, 2000, during the 42nd Grammy Awards
Function: The image inspired the creation of Google Images

Jennifer Lopez ; Versace ; Google

On February 23, 2000, during the 42nd Grammy Awards, recognizing the best artists in the music industry, all eyes were fixed on Jennifer Lopez. This moment has stayed in the collective memory, not for the actress’s musical talents (nominated for her dance song Let’s Get Loud), but for her attire.

J.Lo’s appearance in this gauzy green dress with its extreme décolleté, plunging below her navel, made a huge splash. The light silk mousseline dress—and more specifically, this dress on J.Lo’s body (at least four other celebrities had previously donned this piece designed by Donatella Versace)—is what pushed Google to create its image search engine, as the legend goes. The company itself perpetuates this story. On its official blog8, we can read how in the space of one night, the image became the most popular query on the young search engine, which was mostly text-oriented at the time. When the team devoted to searches realized that they were incapable of providing the results that people wanted—an image of Jennifer in the dress—they decided to create Google Images.

This story was confirmed by the former director of Google, Eric Schmidt, in 2015, in an article published on Project Syndicate. “When Google was launched, people were amazed that they were able to find out about almost anything by typing just a few words into a computer. The engineering behind it was technically complicated, but what you got was pretty rough: a page of text, broken up by ten blue links. (…) So our co-founders Larry Page and Sergey Brin—like all other successful inventors—kept iterating. They started with images. After all, people wanted more than just text. This first became apparent after the 2000 Grammy Awards, where Jennifer Lopez wore a green dress that, well, caught the world’s attention. At the time, it was the most popular search query we had ever seen. But we had no surefire way of getting users exactly what they wanted: J¬.Lo wearing that dress. Google Image Search was born.”9

Google was only two years old and had very few employees as of February 2000, but in July 2001, Google Images was launched. In 2001, the service counted 250 million indexed images; in 2005, over one billion; and today, over ten billion.

In September 2019, Donatella Versace presented her spring 2020 women’s collection during Milan Fashion Week, inspired by the famous jungle-print green dress that Jennifer Lopez had worn during the Grammys ceremony nearly two decades earlier. At the end of the fashion show, a Google image search appeared on giant screens, as the multinational had partnered with the event. Donatella Versace asked the Okay Google assistant to “show me images of the Versace jungle dress,” the nickname given to the dress. The assistant’s female voice obliged: a myriad of dematerialized images of the original 2000 dress were displayed on giant screens. The designer went on: “Okay Google, now, show me the real jungle dress,” and out swaggered J.Lo in the flesh, on the catwalk in a new version of the dress with the same plunging décolleté—images immediately captured by a forest of raised smartphones, and instantaneously bombarding networks, ready to “break the Internet” a second time through the flood of queries.

Marie Lechner
Translation: Sara Heft

The Supermarket of images