A Face Is But a Shape

Grid of shapes rendered in Rhino/Grasshopper and fabricated using a laser cutter.

Software:

• Rhino 7
• Grasshopper
• Adobe Illustrator

Materials

• Cardboard
• Laser Cutter

My project is inspired by abstract representations of faces. I wanted to recreate the basic shapes that novice art students use when learning to draw facial features. I was also interested in discovering how few features—or shapes—are needed for the brain to recognize a face.

Using the drawing guide for reference, I used Python script in Grasshopper to generate a nose (center triangle), eye (left ellipse), smile (bottom triangle) and chin (bottom ellipse). I also included some of the sketch lines that novice illustrators use, for added dimension.

Although it is not pictured here, I began by manually inputting the exact coordinates and size of the shapes. Once I was able to see the "face" the way I originally intended, I replaced hard-coded values with variables. This was important so that I would have more flexibility with sizing and placement while maintaining the same ratio.

Using the blocks in Grasshopper, users can designate an 'x' value that controls the size of the face.

I then sequentially generate the faces in a grid while storing them in a list.

Using a slider in Grasshopper, users can adjust the size of the grid using the y variable. You will also notice that I converted the y value from a float to an int in the code. This is because, even though I changed the slider to only output integers, Python continued to receive a float. I am unsure how to solve this issue, so if you have any suggestions, I would be happy to hear them.

The grid felt a bit repetitive and I decided to rotate various faces to add some variety. In doing this, I also discovered the various images hidden in what I originally intended to be a face. Some people have told me they see a mouse, a sailboat and a zoomed-in image of a cornea. The possibilities are endless.

Other consideration:

I decided against a connected grid. As opposed to my original experimentation discovering the minimum number of shapes needed to recognize a face, I was now interested in the way that result changes when those very pieces were missing (or cut away by a laser.) Would I still see faces in between the holes of the cardboard? What about those sailboats and mice?

After finalizing the project in Illustrator, I let the laser cutter work its magic.

Moments of realization:

1. Although I properly scaled my digital workspace when designing the project, it soon became clear that I did not fully comprehend how small each face was.
2. Because of the size, the ear (oval) and nose (triangle) pieces were cut out as one continuous shape. I previously checked that the lines were not overlapping in Rhino, but it seems that they were too close to individually distinguish between at this scale.
3. Also due to the scale and apparent intricacy of the design, I decided to end the cut about 20 minutes in. Not only was it beginning to run beyond my designated time slot, but I also felt like I had a better understanding of how the machine tackles a cut and how I would adjust my design in the future.
4. Plus, I kind of liked the unexpected variety in the design's pattern. I mentioned earlier that I was not the biggest fan of the monotonous repetition of the design, so stopping the cut earlier provided some welcomed diversity to the texture.

On the final output:

• As mentioned above, I prefer the non-uniformity of the 'half-complete' product.
• Without a distinction between the "ear and nose," the 'faces' lean farther into the 'sailboat' imagery. I also think they continue to look like a mouse with cartoonishly large ears facing the side.
• If we stick to the ocean imagery, the top left quadrant reminds me of sharks swimming in the water. And perhaps the triangles in the upper right quadrant are birds pointed down toward the ocean hoping to catch a fish for lunch.
• I find it really interesting that the laser cutter refused to cut in any easily identifiable order. It started by handling the simple lines in the image and then tackled the bottom left quadrant with ease. It then sporadically began to finish the right side of the board, skipping some columns and rows altogether. I'm not entirely sure why this is.

On future considerations:

• If I were to do this again with my original idea in mind, I would definitely increase the size of the individual faces.
• To avoid the shapes bleeding into each other, I would increase the distance between them as well.
• I might also find it helpful to sketch the dimensions of the sample face on a sheet of paper to reference while I'm working since I better comprehend these things when I can see them in the real world.