Check out this new local ComfyUI workflow to explore 2D images and generate new camera views using Gaussian Splats and Qwen Edit. You generate a Guassian Splat from a single image, then use a LoRA trained on Splats to fill in the missing and distored areas. LoRA and workflow (New node.Json) is the workflow file here: https://huggingface.co/dx8152/Qwen-Image-Edit-2511-Gaussian-Splash Required Viewer https://github.com/CarlMarkswx/comfyui-GaussianViewer I modified the prompt, and improved the viewer so it has an "Auto" scale feature, but I dont know how to update someone's repo on Github. You're mostly limited to the FOV of the original image, but it does perform as welll as the Multi-Camera LoRA in inferring the missing areas. I could put my version on my website for download, but is that bad manners to put a fork of a repository on your own website? I'm also working on depth of field focus and other features.