Powerful ComfyUI Img2img Get Perfect Identity Pose with Auto Prompt

*ComfyUI Advanced Img2img Workflow: Auto-Prompt, Identity & Pose Control (Florence-2, InfuseNet)*

Hey everyone!

In this video, we're diving deep into a powerful ComfyUI workflow that takes your Img2img generations to the next level! Learn how to transform an existing image while gaining precise control over the output's content, identity, and pose.

This workflow demonstrates how to combine several advanced techniques:

✅ *Automatic Prompting:* Uses the Florence-2 vision AI to analyze your base image and automatically generate a detailed text prompt. No more staring at a blank prompt box!
✅ *Identity Control:* Integrates ID Embedding technology to transfer the facial identity from a reference image onto your generated character.
✅ *Pose/Composition Control:* Utilizes a Control Image and InfuseNet to guide the structure and pose of the output image.
✅ *InfuseNet Integration:* Leverages the InfuseNet model for superior blending of identity and pose controls.
✅ *Flux Compatibility:* Designed to work seamlessly with Flux models and guidance.

*How it Works:*

The workflow takes three key input images:
1. *Base Image:* The image you want to transform (used for Img2img).
2. *ID Image:* Contains the face whose identity you want to use.
3. *Control Image:* Provides the desired pose or structural reference.

Florence-2 generates the initial prompt from the base image. Specialized nodes extract ID data from the ID image and pose data from the Control Image. InfuseNet then intelligently combines the prompt, ID embedding, and pose information to condition the KSampler, resulting in a generated image that respects all these controls.

*Custom Nodes Used in This Workflow:*

This workflow relies on specific custom nodes. Make sure you have these installed in your ComfyUI setup. You can install most custom nodes easily using the ComfyUI Manager if you have it installed.

*ComfyUI-Florence2:* For the Florence-2 automatic captioning.
GitHub Link: https://github.com/kijai/ComfyUI-Florence2
*ComfyUI_InfiniteYou:* For the ID Embedding and InfuseNet capabilities.
GitHub Link: https://github.com/bytedance/ComfyUI_Infin...
*rgthree-comfy:* Includes the Power Lora Loader and other useful nodes.
GitHub Link: https://github.com/rgthree/rgthree-comfy

*Get the Workflow (FREEBIE!):*

Want to try this exact workflow yourself?
You can download the workflow JSON file as a *FREEBIE* on my Patreon page!

🎁 *Download Here: patreon.com/DomDom13*

(Note: While the workflow file is free, accessing it on Patreon might require creating a free Patreon account. Remember to place your models, VAEs, LoRAs, and InfuseNet/ID models in their correct ComfyUI directories)

If you found this video helpful or learned something new, please give it a LIKE!
Subscribe to the channel for more ComfyUI tutorials and AI art explorations.
Have questions? Leave them in the comments below!

Thanks for watching!

#ComfyUI #StableDiffusion #AIGeneration #Img2img #Florence2 #InfuseNet #AIArt #Workflow #Tutorial #IdentityControl #PoseControl #TextToImage #ComfyUIWorkflow #AIWorkflow #GenerativeAI #rgthree