Most articles and solutions deal with generating something from nothing. That is, they are focused on the common capability known as ‘text to image’ where the solution can create images based upon the prompt(s) you give it.

There are a lot of nuances to this capability to make sure it can generate images appropriate to your brand but ironically this is one of the simplest capabilities available today.

If you are like many companies, you have invested a lot of time and money in traditional 2D design using a tool such as Adobe Illustrator. Not only do you have a massive number of drawings in this format, but this is also likely the tool that your teams are familiar and very proficient with.

The challenge with these 2D drawings? Well, it’s prettty obvious –they cannot convey what the product will really look like on a person. No matter how much time you spend refining these images, it’s never quite the same as seeing that same thing either re-drawn in 3D or receiving that first physical sample! Both of these approaches of course take time that you don’t always have.

Enter AI! A trained model that understands your brand can now take advantage of that 2D sketch and turn it into photo-realistic images.

For example, we can start with this very simple sketch:

and with a few simple prompts, we can see what this should could look like:

A t-shirt is the most basic example but as you can see, the ability to visualize a photo-realistic rendering within a few seconds is incredible especially if the model is properly trained to your style and fit guidelines!