Stable Diffusion ComfyUI Face Parsing Fine Tune Faces For AI Images
Вставка
- Опубліковано 2 жов 2024
- In these videos, we dive into the fascinating world of face detection and face segmentation techniques for AI images. Join us as we explore how to achieve better performance and refine character faces using the innovative technique called face parsing.
All Resources About This Tutorial : / 105919213
Face parsing utilizes YOLO face segmentation models to detect and label various facial features, including hair, eyes, nose, mouth, and more. This technique is especially effective for close-up shots in AI images, providing detailed and accurate segmentations.
In this video, we demonstrate the usage of ComfyUI, a powerful tool that integrates face parsing into your AI workflows. We walk you through the process of downloading the necessary face parsing models and YOLOv8M model, enabling you to refine and enhance the face and other features in your AI-generated images.
Through a demo workflow, we showcase how to smooth skin, fix wrinkles, and refine specific facial elements using ComfyUI's custom nodes. We also provide tips on adjusting colorations, contrast, brightness, saturation, and hue to achieve the desired results.
Join us in this video to discover the power of face parsing and learn how to integrate it into your AI workflows for impressive face presentations in images and video results. Don't miss out on the opportunity to refine your AI projects and achieve stunning facial details effortlessly.
Have a fantastic day! See you in the video.
If You Like tutorial like this, You Can Support Our Work In Patreon:
/ aifuturetech
Discord : / discord
Custom Node Github github.com/Ryuukeisyou/comfyui_face_parsing
This one is cool. I can fix my old images
Yup 👍
how do i do a pip install for the requirements?
would be super cool if you could prompt each section individually
like
eyes "large, blue, detailed eyes, long eyelashes, almond, upturned"
nose "small, cute, upturned nose"
mouth "open mouth, laughing, teeth, red lips"
chin "strong angular jawline"
hair "short blond hair"
etc... all in one box, mask and prompt
Maybe you can try this, load image, pass to this Segment mask group, get the mask area , use Text condition for the parts that you want to edit. Lastly, send KSampler
another great vida kuya benji
Glad it helps
works on anime faces?
For anime style human face, yup.
@@TheFutureThinker thanks a lot future thinker :D!!
No problem, good tool to try