Skip to main content

Optical Illusions with Ambiguous Objects

Introduction

If you want to make somebody’s brain hurt, a good bet is to show them an optical illusion. These little tricks are fun to show off and a great way to challenge your mind. The only way to increase the fun would be to use mathematics to create your own optical illusion! This can be done through structures called ambiguous objects, which take on a different shape depending on the angle from which you look at them.

Math of Ambiguous Objects

The objective of an ambiguous object is that an individual standing on one side of the object will see a certain shape, while the person standing on the other side of the object will see a completely different shape. How do we make this work? First, let’s decide which shapes we want the observers to see, and then we’ll graph them in the same coordinate plane:

The first shape will be a star, which is parametrized by the following equations: \[ x(t) = 4 \sin(t) + \cos(4t) \] \[ y(t) = 4 \sin(t) - \sin(4t) \] You can also see that our second shape will be the curve that resembles a bat (you may need to use your imagination a little to fully appreciate the bat). It is parameterized by the following equations: \[ x(t) = 4 \sin(t) + \cos(4t) \] \[ y(t) = 10 \sin(t) - \sin(3t) \]

Now, imagine our observers are standing on opposite sides of these curves. We assume that they are looking down on the object at a 45-degree angle:

Let’s say that we want our red friend (let’s call her Sue) to see the bat and for our blue friend (let’s call him Bob) to see the star. If we were to draw a line connecting Sue’s eyes to any point on the bat, we know that every point on that line would appear to be in the same spot from Sue’s perspective. If we do the same for Bob with the corresponding point on the star, we can graph Bob and Sue’s line of sight:

Notice that these two lines intersect at a certain point. This is the point that Sue will perceive as being a point on the bat and that Bob will perceive as being a point on the star. Let’s calculate this mathematically. Say that the vector representing Bob’s eyes is \( \langle 0, 1, 1 \rangle \) and the vector representing Sue’s eyes is \( \langle 0 -1, 1 \rangle \).

Notice that the parametrization for the \( x \) coordinate is the same for both the bat and the star (this will make our math much easier). We will denote this by \( x(t) \). Let’s represent the \( y \) parameterization of our star by \( f(t) \) and the \( y \) parametrization of the bat by \( g(t) \). Then, a coordinate on the star and bat respectively can be represented by: \[ (x(t), f(t), 0) \] \[ (x(t), g(t), 0) \]

In order to generate the lines we saw in the previous image that represented Bob and Sue’s line of sight, we want to find the line that passes through the coordinates above and is also parallel to our observer’s line of sight. These lines are given below, where \( S \) represents Sue and \( B \) represents Bob: \[ r_{B}(q) = (x(t), f(t)+q, q) \] \[ r_{S}(s) = (x(t), g(t)+s, s)\]

Next, in order to find the intersection, we set these equations equal: \[ r_{B}(q) = r_{S}(s) \] \[ (x(t), f(t)+q, q) = (x(t), g(t)+s, s)\] Notice that the \( x \) coordinates are already the same. By setting the \( z \) coordinates equal we can substitute \( q \) for \( s \), and simplify the equation for the \( y \) coordinate: \[ f(t)+q = g(t)+s \] \[ f(t)+s = g(t)+s \] \[ s = \frac{1}{2}\ (g(t)-f(t))\] Plugging this in gives us the resulting coordinate of intersection: \[ (x(t), \frac{1}{2}\ (g(t)+f(t)), \frac{1}{2}\ (g(t)-f(t))\]

When we do this calculation for each point on the original curves, we end up with the following composite:

If we give it some depth, we have a structure that can be 3D printed and placed on the table. It is shown from the side below:

From this angle our object just looks like a weird star-bat hybrid. But when we look at it from Bob’s perspective, he sees the following (he can also see Sue waving from the other end of the table!):

What Sue sees when she looks at the object is the following (she can also see Bob waving from the other end of the table!):

So, with that bit of math we have created an object that takes on two very different shapes when looked at from opposite angles.

Why this example?

I chose these shapes because, first of all, I think the star parametrization is very cool looking and I’ve used it in a lot of previous assignments, so I wanted to keep the trend going. Secondly, I thought these two shapes were different enough for the illusion to be mind-boggling, since the bat shape is much wider and less pointy than the star. Both shapes have a lot of curves, so it’s hard to imagine at first how one object could appear to be both. I also wanted the shapes I used to be easily recognizable. Thirdly, these shapes worked out well logistically since they have the same parametrization for the \( x \) coordinate.

Author: Sarah Bombrys

Comments

Post a Comment

Popular posts from this blog

Do Over: Integration Over a Region in a Plane

Throughout the semester we have covered a variety of topics and how their mathematical orientation applies to real world scenarios. One topic we discussed, and I would like to revisit, is integration over a region in a plane which involves calculating a double integral. Integrating functions of two variables allows us to calculate the volume under the function in a 3D space. You can see a more in depth description and my previous example in my blog post, https://ukyma391.blogspot.com/2021/09/integration-for-over-regions-in-plane_27.html . I want to revisit this topic because in my previous attempt my volume calculations were incorrect, and my print lacked structural stability. I believed this print and calculation was the topic I could most improve on and wanted to give it another chance. What needed Improvement? The function used previously was f(x) = cos(xy) bounded on [-3,3] x [-1,3]. After solving for the estimated and actual volume, it was difficult to represent in a print...
Post a Comment
Read more

Minimal Surfaces

Minimum surfaces can be described in many equivalent ways. Today, we are going to focus on minimum surfaces by defining it using curvature. A surface is a minimum surface if and only if the mean curvature at every point is zero. This means that every point on the surface is a saddle point with equal and opposite curvature allowing the smallest surface area possible to form. Curvature helps define a minimal surface by looking at the normal vector. For a surface in R 3 , there is a tangent plane at each point. At each point in the surface, there is a normal vector perpendicular to the tangent plane. Then, we can intersect any plane that contains the normal vector with the surface to get a curve. Therefore, the mean curvature of a surface is defined by the following equation. Where theta is an angle from a starting plane that contains the normal vector. For this week’s project, we will be demonstrating minimum surfaces with a frame and soap bubbles! How It Works Minimum surfac...
Post a Comment
Read more

Do Over: Ruled Surfaces

Why to choose this project to repeat For the do over project, I would like to choose the ruled surfaces. I don't think my last project was creative, and the 3D printed effect was not very satisfactory. In the previous attempts, all the lines are connected between a straight line and a circle. This connection structure is relatively uncomplicated. The printed model has too many lines, resulting in too dense line arrangement. The gaps between lines are too small, and the final effect is that all the lines are connected into a curved surface, which is far from the effect I expected. What to be improved In this do over project, I would like to improve in two aspects. Firstly, a different ruled surface is chosen. In the previous model, one curve is a unit circle on the \(x-y\) plane, and the ruled surface is a right circular conoid. In this do over project, it is replaced by two border lines. Each borderline is in the shape of an isosceles right triangl...
Post a Comment
Read more