Posted by s.hettrick on 25 March 2014 - 10:00am

By Hannah Dee, computer science lecturer, Aberystwyth University.

When I was approached to write a guest post on women in software, my first thought was to try and pull together another post about the leaky pipeline, school science, or girls toys. But that’s not the field in which I do most of my software development. It’s what I tend to pontificate on, but not what I research. I’m a vision researcher. So, could I come up with a computer vision topic that was somehow gendered? Easy!

When doing research in computer vision or image processing, it's useful to have a test image or two. Writing programs that reduce noise, alter brightness, or enhance edges is all very well and good, but without test images, we can't know if they work. Early on in vision science, the acquisition of images was hard, and there were a handful of images everyone used. This was partly due to expediency (not everyone had access to a scanner) and partly due to comparability (we want to be able to see the results of each algorithm on the same image or set of images). Today, nearly everyone has a digital camera as part of the device in their pocket, in the 70s and 80s such devices simply didn't exist.

At the very beginning of the discipline that's now become computer vision, sometime in the early 70s - probably early 1973 - a researcher was looking for a test image. Alexander Sawchuk (now a professor at the University of Southern California) scanned one for this researcher: it was the centrefold of a Playboy magazine, and that scan has now entered into our scientific culture in a way the originators could never have imagined. The woman is wearing a floppy hat, and is gazing over her bare shoulder towards the camera. As a 512 pixel squared image, the woman's body is cropped at the shoulders; the suggestion of nudity is there, but the image itself is decidedly safe for work. Her name is Lena Söderberg. The researcher’s name is lost to internet history, as is the name of the person who actually brought the porn to work.

Lena (sometimes spelled Lenna, which is the anglicisation of her name used by Playboy) has been called “The First Lady of the Internet”. Her image has been printed on a chip, shrunk to the size of a human hair, blurred, sharpened, and undoubtedly enhanced. She appeared on the cover of the journal Optical Engineering in 1991, and the model herself was invited to a conference of the Image Science and Technology society in 1997. Her picture is on the wall, tastefully framed, in at least one reputable computer vision lab. For computer vision researchers, her image is everywhere.

In defence of the use of Lena, Hutchinson said “... the image mixes areas of light and dark, fuzzy and sharp, detailed and flat—providing a stiff test for an image processing algorithm”. (I am unsure whether the word stiff in that sentence is a deliberate double-entendre. In a sense, I hope it is: if we're going to have scholarly articles about the use of pornography in our science, let them at least try to be funny.) Interestingly, Hutchinson's article talks about issues with Lena, from the copyright (Playboy were, for a time, not best pleased by the widespread distribution of one of their images) to the equalities issues. Many journal editors, whilst uncomfortable with the idea of such images appearing in their journals, were even less comfortable with the idea of banning such images.

But, as Hutchinson's article was published in 2001, we can surely assume that things have changed – that was over ten years ago. Our discipline is becoming more aware of the issues of minorities: the fact that women in technology can feel isolated is something that everyone working in the field probably knows by now. The usage of Lena must have gone down. The controversial nature of the image (there are discussion threads on the topic stretching back years, on some social media platforms) must make people think twice. Surely people realise that women, as a minority in the field, might be put off by statements like “... 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.” [1]

Right?

Wrong. At ISISPA2013, the IEEE International Symposium on Image and Signal Processing and Analysis, during a single day of the talks I saw her six times. She's included in the standard OpenCV download (at least twice – as a test image in the both the c++ and python directories). On the machine I'm using to type this blog post, despite never having used Lena in a piece of research, and never having deliberately downloaded her, I've just done a search, and I'm amazed. I've got 18 copies (excluding cached thumbnails and the directory of files I've gathered whilst researching this article). It seems that whenever you download a piece of vision code, Lena still comes along for the ride.

As a vision researcher, I'm used to being the odd one out: at the first workshop I ever spoke at I was the only woman in the room. Generally, that doesn't bother me. I'm not entirely sure what I think about Lena though – having decided she would be a good topic for an article, I read up on the history, and to be honest, I still find it a bit bizarre. But despite my avowedly feminist stance, I’m somehow unable to get that annoyed about it.

The fact that there's a historic Playboy image at pretty much every conference I go to, and on the walls of my colleague's labs, and downloaded with every single image processing library I use, well... on the one hand, it's part of that drip-drip-drip of strangeness that comes from working in a male-dominated field, where the topics of conversation and the general attitude can be a little disconcerting. But on the other hand, with changing cultural attitudes, and the effect the internet has had on pornography, the entire centrefold (yes, you can easily find it online if you look) seems very tame indeed by today's standards. And the crop that is used in image processing research is, well... I've developed quite an affection for the picture. It's one of the quirks of computer science. So when I was asked what picture we should use to illustrate this blog post, there was only one choice.

Of course, we really had to use a Lena image on this page (although subject to Gaussian Blur with a 60px by 60px kernel). You can probably still see that it’s just a shoulder and a head in a floppy hat: what’s all the fuss about? Well, we all know what it represents and we all know where it comes from. And sometimes, there’s more to an image than meets the eye.

[1]: David C. Munson, Junior. “A note on Lena”, 1996 IEEE Transactions on Image Processing, vol 5 issue 1.