ComfyUI Node: LivePortrait

LivePortrait:

The LivePortraitNode is designed to create dynamic and realistic animated portraits from static images. This node leverages advanced machine learning models to analyze and animate facial features, providing a seamless and lifelike transformation. By utilizing pre-trained models and sophisticated algorithms, the LivePortraitNode can generate animations that mimic natural facial movements, making it an invaluable tool for AI artists looking to bring their portraits to life. The node's primary goal is to simplify the process of creating animated portraits, allowing you to focus on the creative aspects without worrying about the underlying technical complexities.

LivePortrait Input Parameters:

dsize

This parameter defines the desired size of the output image. It is crucial for ensuring that the generated portrait fits the required dimensions for your project. Adjusting this parameter will directly impact the resolution and aspect ratio of the animated portrait. There are no specific minimum or maximum values provided, but it should be set according to the needs of your project.

scale

The scale parameter controls the scaling factor applied to the image. This can be used to zoom in or out of the portrait, affecting how much of the face is visible in the final animation. Adjusting the scale can help emphasize certain facial features or fit the portrait within a specific frame. The default value is typically set to 1.0, representing no scaling.

vx_ratio

This parameter adjusts the horizontal position of the portrait within the frame. By modifying the vx_ratio, you can shift the portrait left or right, which can be useful for aligning the face within a specific area of the output image. The default value is 0.0, meaning no horizontal shift.

vy_ratio

Similar to vx_ratio, the vy_ratio parameter controls the vertical position of the portrait. Adjusting this parameter allows you to move the portrait up or down within the frame. This is useful for ensuring that the face is centered or positioned according to your artistic vision. The default value is -0.125, providing a slight upward shift.

LivePortrait Output Parameters:

v_path

This output parameter provides the path to the generated video file containing the animated portrait. It is essential for accessing and utilizing the final animation in your projects. The v_path will point to the location where the video is saved, allowing you to easily retrieve and view the animated portrait.

output_path_concat

This parameter gives the path to the concatenated output, which may include additional processing or combined results from multiple steps in the animation pipeline. It is useful for accessing the final, polished version of the animated portrait, ensuring that all necessary adjustments and enhancements are included.

LivePortrait Usage Tips:

• Ensure that the input image is of high quality and well-lit to achieve the best results in the animated portrait.
• Experiment with the scale, vx_ratio, and vy_ratio parameters to find the optimal framing and positioning for your portrait.
• Utilize the output paths (v_path and output_path_concat) to review and refine the generated animations, making any necessary adjustments to improve the final result.

LivePortrait Common Errors and Solutions:

Error: "Model file not found: landmark.onnx"

• Explanation: This error occurs when the required model file for facial landmark detection is missing.
• Solution: Ensure that the landmark.onnx file is present in the specified directory (liveportrait_model). Verify the file path and download the model if necessary.

Error: "Failed to load source image"

• Explanation: This error indicates that the source image could not be loaded, possibly due to an incorrect file path or unsupported image format.
• Solution: Check the file path of the source image and ensure it is correct. Verify that the image is in a supported format (e.g., JPEG, PNG).

Error: "Inference configuration missing"

• Explanation: This error suggests that the inference configuration file (models.yaml) is not found or improperly set up.
• Solution: Verify the presence and correctness of the models.yaml file in the specified directory (current_directory/LivePortrait/src/config). Ensure that all paths and settings within the file are correctly configured.

Error: "Invalid crop configuration"

• Explanation: This error occurs when the crop configuration is not properly set, leading to issues in processing the input image.
• Solution: Review the crop configuration settings and ensure they are correctly defined. Check the crop_cfg parameter and adjust as needed to match the input image dimensions and requirements.