Hey there! I’m excited to finally share something I’ve been working on — FacePop, my brand-new extension for AUTOMATIC1111’s Stable Diffusion Web UI. If you’ve been looking for a way to take control of face detection and enhancement in your images, this tool might just be what you’re looking for. I created it as a better solution to some of the limitations in the current tools, like Zoom Enhancer, and it’s packed with cool features that give you way more control over facial enhancement in your images.

So what is FacePop exactly, and why did I create it? Well, if you’ve ever worked with Stable Diffusion, you know that while the image generation is great, sometimes the faces just don’t come out quite right. This is where FacePop comes in. It’s designed to detect faces in an image, crop them, upscale them, enhance them with some nifty processing tricks, and then seamlessly blend them back into the main image. It’s like giving the faces in your images a personal makeover!

Why I Built FacePop

Initially, I used Zoom Enhancer to zoom in on faces and fix the quality, but I found myself wanting more control. I also wanted to integrate with some other great plugins like ControlNet, ReActor, and After Detailer. FacePop lets you do all that and more, without needing to rely on other extensions like Unprompted.

Basically, I wanted a way to get faces in my images looking as good as possible, but I didn’t want to jump through hoops or have a ton of separate tools cluttering up my workflow. With FacePop, everything’s built-in and ready to go.

How Does FacePop Work?

Here’s a quick breakdown of what FacePop does:

1. Detect Faces: It accurately detects faces in the image using Mediapipe. Once the faces are found, it identifies key facial landmarks to make sure everything is aligned.
2. Crop & Upscale: The tool crops out the faces, scales them up (with padding if you want), and gets them ready for processing.
3. Enhance the Faces: Each face gets processed separately — you can do things like color correction, sharpening, background removal using MODNet, and more.
4. Mask & Blend: After processing, it creates a mask around the faces and blends them back into the original image. This means no awkward edges or mismatches.
5. Final Touches: Once the faces are placed back in, the whole image gets processed again for any final tweaks or adjustments.

The best part? You can fine-tune everything! From face width and height to padding and detection confidence, it’s all in your hands.

Why Should You Care?

Whether you’re a digital artist, photographer, or just someone who loves generating images, you’ve probably run into situations where the faces in your artwork just didn’t look quite right. FacePop fixes that. You don’t have to manually touch up every image anymore — just let the tool handle it.

It’s also integrated with ControlNet for advanced manipulation, and it’s designed to play nicely with popular tools like ReActor and After Detailer, giving you the flexibility to enhance faces while still maintaining creative control.

Easy to Use

FacePop is super easy to install and use. If you’ve got the Stable Diffusion Web UI set up already, installing this extension is a breeze. Just grab it from the Extensions tab using the URL:

https://github.com/TheCodeSlinger/FacePop.git

Once it’s installed, you can access all its features directly in the Img2Img interface, where you’ll find a ton of options for tweaking and customizing how faces are processed.

What’s Next?

I’m planning to keep improving FacePop and add more features based on user feedback. So if you try it out and have ideas for improvements, hit me up! There’s a lot of potential to keep pushing this tool further, especially as more people start using it in different workflows.

So, if you’re tired of low-quality faces in your Stable Diffusion images or just want more control over how faces are enhanced, FacePop might be the tool you’ve been waiting for. Download it, give it a spin, and let me know what you think!

That’s it for now! I’m super excited to see what you all do with FacePop and how it fits into your creative process. Stay tuned for updates and feel free to share your results or any feedback. Happy generating!

Cheers!