A Playboy Based Test Target has No Place In Imaging Science

I recently attended, and volunteered to help facilitate, a short course at the Imaging Science and Technology conference “Electronic Imaging” where the very famous Lena test image was used to demonstrate various image compression concepts. In its first usage in the class, there was at least some mention of the historical context. Lena’s portrait, shot for the centerfold of a 1972 issue of playboy magazine, has been used as test material in a huge number of influential or foundational papers in the field of imaging science and perceptual imaging. However, her image was used a further two times in the class with no compelling scientific reason to demonstrate concepts about texture compression in images. Even after these usages, other examples were provided demonstrating the same algorithm, making the usage of Lena’s portrait simply gratuitous. It’s long past time for her to retire from imaging science, as Lena Forsén, the original model, suggested in 2019. [https://www.losinglena.com/]

I’m conflicted in writing this today because the course instructors are truly world class experts, and, as I will describe later, both voiced support for losing Lena during a Q&A session. The conference is organized by IS&T which is a women led organization and already has a moratorium in place regarding the usage of Lena. Dr. Susan Farnand, the current president of IS&T has already reached out to me to say that IS&T fully supports the removal of Lena from imaging science work. The IS&T along with Nature and SPIE each banned the usage of the Lena image in journal papers and classes in 2019. Dr. Farnand committed to reminding all IS&T editors and instructors of this policy and I appreciate her work. It’s clear to me that Dr. Farnand and the IS&T in general are on the right side of this issue, and so I don’t want to sit and lambast them for some material that mistakenly made it through review.

Still, the relevant bans by academic organizations have only been in place for a few years. And it’s clear that some researchers see no issue with the usage of a playboy-centerfold as a test target in their research, as evidenced by IEEE’s continued support of using Lena test images. I feel compelled to bring this discussion into the open and to say stop using lena now.

After the third and unnecessary usage of the Lena image in the class I posted my question to the chat box. “What is the purpose of using the Lena image in the previous slides? My understanding is that it’s usage will result in paper rejections from Nature, SPIE, and IS&T.” My question sat there for some time while the instructors moved forward with the material. After the end of the material the instructors addressed my question live, and their full responses are in the class recording. Genuinely, I found their responses very enlightening about the usage of the image both historically and today.

First one instructor raised the quick point that (paraphrasing mine) old habits die hard. Which is, I think, to say that the image is used in a lot if historical material and for many researchers it is default. Due to its historical inertia many reviewers or researchers will practically expect the Lena image to be used in order to draw comparison to older papers and methods. She (the instructor) was not using this to voice support for its use, but rather providing background. She then added that, especially in the context of studying older research there was often no alternative. There were very few digital images with known visual characteristics available, which image perception research requires. Even in the case that an original author, like herself, wanted to use other material there was often no suitable alternative available. That is not the case today and she voiced support of the policies in place by SPIE and IS&T.

Lastly, she also commented that it’s long past time for the IEEE to follow suit with the other leaders in this field and ban the usage of Lena as a test image in it’s journals. More on that later.

After she spoke about the historical context of Lena, the other instructor politely added his thoughts. The usage of the Lena image, aside from it’s problematic history, is a terrible choice for technical reasons. In the case that you are studying perceptually lossless and high quality image processing algorithms using a low resolution scan of an image printed in half tone (partial resolution and partial color depth) on magazine paper is a really bad basis for your research. Developing high quality compression algorithms requires image comparison of the compressed image to high uncompressed quality images, which disqualifies the usage of the Lena headshot. He also apologized and voiced his support for the standing moratoriums at IS&T, Nature, SPIE and called on IEEE, at least in that class, to follow suit.

I genuinely believe these instructors are sincere in their support for #LosingLena. I think this material in their slides is only present due to historical pressure by other researchers, and out of date slides that were produced prior to 2019. Still, their response felt at the time very much like them saying “do as I say not as I do.”

A call to action: The leaders and veteran researchers in our field of imaging science must renew their commitment to #LosingLena and lead by example. For professors, instructors, and veteran researchers it’s time for you to review old material that you might be presenting this year and make every effort to update it without using the Lena headshot. It’s time for the IEEE to make the same pledge that other academic organizations made 3 years ago. And it’s time for Nature, SPIE, and IS&T to review the pledges they made in 2019 and ensure that their editorial policies are being followed.

Lead by example.

In the course of preparing these comments I spoke with another professor of mine at Rochester Institute of Technology who is teaching a class on color imaging this semester. He commented that the usage of Lena’s image is present in many old textbooks and that he encountered several usages in different readings that he was considering for the graduate course this year. He asked, “Should bring to the attention of the class or not? Is it a useful cautionary tale, or is bringing it up simply perpetuating a misogynistic inside joke?”

I think that is a question on the minds of many researchers currently. Is this topic something that I should bring up? Will it just draw more attention to and therefore lengthen the tenure of the misogyny in our field? I’d much rather spend my time quashing this problem.

It is wrong to ignore history, and it’s used in many important studies. If you are talking about the historic context of the advances in perceptual imaging it might be relevant. Even in those cases, unless you are discussing the impacts of the usage of the playboy-based test target in a male dominated field it seems gratuitous to include the image. Names, papers, citations, re-implementations of algorithms are all much better tools for talking about the advances in the field and don’t require its use. It’s only been banned from several journals for a few years and while that should have been plenty of time for people to take notice. It can also be expected that Lena will hang on for a little longer and examples will make it through. For advanced students, I think learning and reconciling the role of misogyny in the field is valuable.

But for undergraduate students outside of an ethics specific course, and for highschool students, the depiction of a nude portrait from a playboy magazine must end now. “But it’s just a headshot” some people will say, but that doesn’t matter. Listen to the story of Maddie Zug who recounted her experience as a 16 year old girl in a computer science classroom in Virginia. Zug commented on her experiences in the 2019 documentary Losing Lena as well, “It made me feel like… women were made to be this joke in the context of this class where I was supposed to be learning as an equal.”

Bringing up the ban many years from now will probably be gratuitous and perpetuating, and maybe it already is, but I don’t think we are there yet.

There’s also a point to be made about Lena’s feeling about her own modeling work. Lena Forsén was interviewed, briefly, in 2019 by UCB Rhetoric PhD Candidate Linda Kinstler for Wired Magazine. In the article, Kinstler writes that Lena is and was quite proud of her work. And in fact, she contributed much more to imaging science than people even realize. Some time after the Playboy photoshoot Lena moved to Rochester, NY where she worked as a model for Kodak. Her images from Kodak have been used in many important research papers on the perception of image quality throughout the 80s and 90s, they were used in the various calibration processes for the production and development of new film stocks, and she posed in images for Kodak’s marketing materials.

And of course, she should be proud of that work. And, unlike many women in the early history of technology, she’s received a great amount of modern recognition for those contributions. She has been invited to and presented awards at at least two large imaging science conferences. Lena’s only regret from the 2019 Wired article was that she wished she had been paid more for her contributions and work. Her likeness was used all over the world for decades by Kodak, for which she I doubt she received any residual licensing fees. In addition, Kinstler write that Lena seemed alarmed that her proud photoshoot could be playing a part in hurting or discouraging young women from joining the tech field.

I would encourage all readers of this article to re-read the articles and accounts by Maddie Zug (the above linked WaPo op ed) and by Linda Kinstler (Wired), as well as watch the full documentary about #LosingLena on Facebook Watch.

Thank you for taking the time to read this far, and please, if you have the time read further and listen to the women who are calling out misogyny in the tech and research industries every day. It doesn’t matter that we are not the group or generation of researchers who have caused this problem or led the industry down this path. But we can be the intentional solution.

I’d like to take the time to again thank Dr. Susan Farnand, president of the IS&T, and Suzanne Grinnan, Executive Director, for their work on the Electronic Imaging conference. I’d also like to thank the staff and professors at RIT who I get to thoughtfully engage with on challenging topics such as this. I’m proud that I get to work with such dedicated and thoughtful researchers every day.

I am currently a PhD student at RIT’s Munsell Color Science Lab. Though for the purposes of this article Dr. Farnand reached out on the behalf of IS&T, I’d like to acknowledge that she is also a faculty member at RIT’s Munsell Color Science Lab.

Leave a comment

Leave a Reply