Gemini Omni - Outfit Swap
Upload a video of a person and a reference image of the outfit you want them to wear. Gemini Omni maps the clothing onto the moving subject while preserving their appearance, movement, camera framing, and original environment—ideal for virtual try-ons, fashion concepts, and styling previews.
概要
This ComfyUI tutorial walks you through a focused outfit-swap workflow powered by the GeminiVideoOmni node. You load a person video with LoadVideo (Your Footage) and a single outfit photo with LoadImage (Outfit Reference). The GeminiVideoOmni node takes these inputs plus a plain-language instruction (provided via a PrimitiveNode) to map the outfit from the reference image onto the moving subject in your clip, while preserving identity, motion, camera framing, and the original background. The generated result is written out with SaveVideo.
入力するもの
- ワークフローを読み込み中
ComfyUI で実行できます。 このワークフローはこのページでは実行できません。すべてが読み込まれた状態でComfy Cloudで開くか、ご自身のComfyUI用にダウンロードしてください。
Cloudで試すよくある質問
No. The GeminiVideoOmni node uses the source video as structure guidance and aims to preserve the original scene. A clean subject and fewer occlusions will improve results, but no masking is required.
Use a single, sharp image with the garment clearly visible, minimal wrinkles or motion blur, and a view that matches the pose and angle in your video. Crop away faces and busy backgrounds if possible to reduce ambiguity.
Shorter clips (e.g., under 10–15 seconds) are recommended for faster iteration. Longer or high-resolution videos increase compute and time. If you hit memory or time limits, downscale the video or process in segments.
It can, but complex elements (flowing skirts, transparent fabrics, shiny materials, intricate logos) are harder to track and may flicker. Choosing a clear reference and moderate motion helps improve temporal consistency.