An algorithm that automatically suits clothes?

Question

Imagine that a client asked you to develop a piece of software that would satisfy some pretty crude specifications as follows:

1: it will be sold in the fashion industry

2: users will be people who “design clothes and things” (there is a certain term for this, but it does not occur to you)

3: For various reasons, it would be useful to be able to quickly prototype projects and see how they look on models. To make it more useful, he must present 3D models so that designers can look at him from different angles, if necessary.

4: To be useful, users should be able to design their own outfits using any 3D modeling tool that they like (it can even be built-in), and then import it into this system. Similarly, users should also be able to create their own models.

5: Users will be able to choose the model in which they want to try the equipment. Naturally, people come in different proportions, so they may want to adjust the model accordingly to see how it looks on different people.

6: Consequently, clothing should be adjusted for proportions. It is impossible to create the same outfit again and again just to make minor changes, so this must be done by the computer.

Most of the system sounds pretty simple. But in the end, 3D material appears.

In particular, the most difficult part of the system that I present will be the actual “clothes fitting” algorithm, since you can create static objects and see how they look, but be able to change them on the fly and still have them fit naturally, it seems completely insane.

Are there any other such tools? What tools will facilitate this decision? Has anyone ever worked on something like this?

Answer 1

There is a lot of hardcore computational geometry in it, and while bits and pieces of it exist, as far as I know, no one has done this before.

Oddly enough, clothing can be one of the easiest parts. You would define a fabric constructor with parametrically dimensional components, simulate a model wearing it in a certain size and in different poses, and perform a tightness measure (the number of deformations) in each pair of pose size. Then you define a cost function with a penalty for size and density, then perform gradient descent to minimize it.

But first you need tissue modeling (hard), three-dimensional modeling of people in various poses (hard) and a modeling tool for templates with a parametric shape (hard).