Lenna’s Meaning: A Discussion of a Digital Image

The image of Lenna, probably is the most transmitted and analyzed digital image in the world. So, I think it is a perfect example for discussing the relation between digital information and symbolic meanings.

1

Lenna’s Image

In 1973, in order to complete a research paper in image processing, an assistant professor at USC called Alexander Sawchuk scanned a 5.12X5.12-inch-square of a centerfold from Playboy with three analog-to-digital converters[i]. It became the most widely used standard test image thereafter[ii].

1. The Processing and Transmitting of Information Are Irrelevant to the Meaning

Lenna’s image was selected without specific purposes, demonstrating that digital processing, at least the research of digital processing is irrelevant to the meaning. The reason for Sawchuk to choose this image was that he was tired of those boring pictures existing in his system. Just in time, a colleague came by with an issue of Playboy. Being attracted sexually, of course, he decided to use Lenna’s image on Playboy in his paper. The arbitrariness of the selection showed that the meaning of the image had nothing to do with his research. Even though from a hindsight, it was evident that the image was perfect for image processing algorithm tests because it mixed different properties very well, such as “light and dark, fuzzy and sharp, detailed and flat[i]”, those properties were merely physical attributes of pixels on the screen when it was displayed, also irrelevant to the meaning of the image.

屏幕快照 2016-10-17 下午9.58.34

Some examples of image processing tests using Lenna’s image. Clockwise from top left: Standard Lena; Lena with a Gaussian blur; Lena converted to polar coordinates; Lena’s edges; Lena spherized, concave; Lena spherized, convex. Source: http://www.cs.cmu.edu

The digital representation of the image is also irrelevant to the meaning. Sawchuk used three analog-to-digital converters in his scanner. The converters were responsible for red, green, and blue respectively. That is to say, each pixel in the scanned image is digitally represented by and only by three numbers[iii].

2

Each pixel has three numbers representing red, green, and blue respectively.

As we can see from the website (as shown below) of USC Signal and Image Processing Institute (SIPI) where Sawchuk used to work, the original image consists of 512X512 pixels. Each pixel has three numbers representing three colors. Each number is 8 bits (1 byte), so each pixel is 3 bytes. In turn, the whole image is 3X512X512=786,432bytes, that is 768 kb, as shown in the screenshot below.

4

A screenshot of SIPI website including Lenna’s picture.

So, basically, the image that we see as a naked girl with a blue feather hat is merely made of 786,432 numbers. In other words, in the digital representation of the image, there are only numbers, no naked girl, no hat, no feather, and no symbolic meanings.

2. However, the meaning is reserved, and extended

If there are only numbers, why do people enjoy talking about the story about Lenna? Actually, Lenna was seen as a symbol for the field of image processing, so important in computer science that she was invited to many academic conferences and was surrounded by crazy fans immediately. The sale of Lenna’s issue (Nov. 1972) was over seven million copies, becoming Playboy’s best-selling issue ever.

I think the answer lies in three levels.

First, the meaning of the image is reserved “in the physically observable patterns[iv]” of numbers. According to Paolo Rocchi, “information always has two parts—sign and referent. Meaning is the association between the two[iv].” The associations are stored in our brains. For example, great contrast in brightness is perceived by human as an edge. When the edges form a specific pattern, it would be associated with a face.

5

The red lines show edges that are associated with a human face.

Sometimes, we don’t need a high resemblance to perceive a pattern as a face.

6

A hill on the Mars, misperceived as a human face due to the pattern caused by great contrast in brightness.

After we recognized a human face in Lenna’s image, a higher abstraction—the facial expression, body gesture, and accessories—indicates the gender. In this way, we receive the meaning of Lenna’s image—a naked girl wearing a feather hat.

Second, what is the meaning of “a naked girl wearing a feather hat”? It means sexual attraction to males. What does it mean to use a sexually attractive image in a highly academic context? It might mean a male chauvinistic tendency in the academic community. That was why the usage of Lenna’s image caused controversy. A high school girl even published an article on the Washington Post to discuss the negative impact of Lenna’s image to female students who decided to keep away from Computer Science[v].

Third, as a standard test image, Lenna’s image was frequently associated with computer science. Over time, Lenna became a symbol in computer science. In November of 1972, if people saw this image, they would say: “Oh, she is a Playboy Playmate.” But in October 2016, if people saw this image, they would say: “Oh, this is the famous Lenna. She is somehow important in the history of computer science.”

3. Discussion

Although information transmission is irrelevant to human meaning, the design of information transmission is relevant to the human meaning-making process. The reason why only three converters were used by Sawchuk was based on human perception of colors, which was in turn based on the three types of cone cells in the retina. Each type of cone cells could sense a part of the electromagnetic spectrum that was perceived as red, green, or blue. I’m sure the original image on paper reflects infrared ray, too. But infrared is outside human visible spectrum, so the information in the infrared spectrum is useless, at least in this context. That’s why RGB system was enough to represent and transmit most meanings in images.

At last, I was wondering, computers are able to recognize patterns such as faces and houses, too. The associations are stored in algorithms and memory. Does it mean computers are capable of meaning making, too?


References

[i] Jamie, Hutchinson. 2001. “Culture, Communication, and an Information Age Madonna.” IEEE Professional Communication Society Newsletter 45 (3).

[ii] “Lenna.” 2016. Wikipedia. https://en.wikipedia.org/w/index.php?title=Lenna&oldid=737697952.

[iii] Prasad, Aditya. 2015. “Ideas: Discrete Images and Image Transforms.” Ideas. September 19. http://adityaarpitha.blogspot.com/2015/09/discrete-images-and-image-transforms.html.

[iv] Denning, Peter J., and Tim Bell. 2012. “The Information Paradox.” American Scientist 100 (6): 470–77.

[v] Zug, Maddie. 2015. “A Centerfold Does Not Belong in the Classroom.” The Washington Post, April 24. https://www.washingtonpost.com/opinions/a-playboy-centerfold-does-not-belong-in-tj-classrooms/2015/04/24/76e87fa4-e47a-11e4-81ea-0649268f729e_story.html?utm_term=.059043a988f1.

发表评论

电子邮件地址不会被公开。 必填项已用*标注