Utilizing Stable Diffusion
for Gender Affirming Care
Staying true to my commitment of learning new technologies and staying up to date with emerging trends in digital media and editing, I have recently begun working with Stable Diffusion, a deep learning text-to-image model.
What Is Stable Diffusion?
Originally designed to generate detailed images from text descriptions (text-to-image),
Stable Diffusion has undergone recent updates and releases, offering a range of features that can be utilized within the digital media industry. These updates provide more control and customization options than ever before
How I Use Stable Diffusion
Using the Stable Diffusion Webui, I can take advantage of the inpainting feature. This involves uploading an image generated within Stable Diffusion or any other image and directly drawing a mask onto the areas I want to edit. I then input a text prompt.
The program utilizes the text prompt to generate and replace parts of the image, resulting in an edited version that matches the description I provided.
Where Does Gender Affirmation
Come into Play?
Recognizing the potential of this program, I began experimenting with my own childhood photos to see if I could modify them to align with my gender identity today.
For example, I have been able to remove dresses and replace them with formal attire typically associated with boys. Additionally, I have edited my high school graduation photos to showcase a black cap and gown.
How Does This Help?
By altering and editing old physical memories I used to hide away, I can now proudly display milestones within my own home again without feeling detached or embarrassed with how I looked
during those moments.
The end goal is to have a dependable and efficient method in altering photos of the trans community that doesn't require heavy photoshop skills to clean up the final product.
I hope that other digital media professionals, especially those who identify as LGBTQ+, will also recognize the power of artificial intelligence and utilize it to revolutionize our community's representation by no longer hiding past versions of ourselves and our milestones away.
Before
After
Other stable diffusion images, where I utilized IMG2IMG and a text prompt to change the style / layout of the original photo.
