ComfyUI Node: Face Shape Match

Face Shape Match:

FaceShaper is a powerful node designed to detect and analyze facial landmarks in images, enabling precise face shape matching and manipulation. This node leverages advanced facial recognition models to identify key facial features such as eyes and mouth, and can interpolate between different facial images to create smooth transitions. It is particularly useful for AI artists looking to blend or morph faces in their artwork, ensuring that the facial features align accurately. By using FaceShaper, you can achieve high-quality face alignment and transformation, making it an essential tool for any project involving facial image processing.

Face Shape Match Input Parameters:

analysis_models

This parameter expects a model that can perform facial analysis and interpolation. It is crucial for the node to function as it provides the necessary algorithms to detect and manipulate facial landmarks.

imageFrom

This parameter takes the source image from which the facial features will be extracted. The image should be in a format that can be processed by the node, typically a NumPy array representing the image in RGB format.

imageTo

This parameter takes the target image to which the facial features will be aligned or interpolated. Similar to imageFrom, this should be a NumPy array representing the image in RGB format.

landmarkType

This parameter specifies the type of facial landmarks to be used. Options include 81, 68, and 5, which correspond to different sets of facial landmarks. The choice of landmark type affects the precision and detail of the facial feature detection.

AlignType

This parameter determines the method of alignment for the facial features. Options include "Width", "Height", and "Landmarks". Each option adjusts the alignment based on different criteria, such as maintaining the aspect ratio or aligning specific landmarks.

Face Shape Match Output Parameters:

output

The output is a list containing two images. These images are the result of the interpolation between the source and target images, with facial features aligned according to the specified parameters. The images are returned as PyTorch tensors, ready for further processing or visualization.

Face Shape Match Usage Tips:

• Ensure that the source and target images are of similar size and resolution for optimal results.
• Experiment with different landmarkType values to find the best fit for your specific use case.
• Use the "Landmarks" option in AlignType for the most precise alignment of facial features.

Face Shape Match Common Errors and Solutions:

The 5 point landmark model is not available. Please download it from https://huggingface.co/matt3ounstable/dlib_predictor_recognition/blob/main/shape_predictor_5_face_landmarks.dat

• Explanation: The required 5-point landmark model is missing from your system.
• Solution: Download the model from the provided URL and place it in the appropriate directory.

The face recognition model is not available. Please download it from https://huggingface.co/matt3ounstable/dlib_predictor_recognition/blob/main/dlib_face_recognition_resnet_model_v1.dat

• Explanation: The required face recognition model is missing from your system.
• Solution: Download the model from the provided URL and place it in the appropriate directory.

No faces detected in the image.

• Explanation: The node was unable to detect any faces in the provided image.
• Solution: Ensure that the image contains clear and visible faces. Adjust the image quality or lighting if necessary.