mettaagi/facenet-faces-vectors-dataset

Face Vectors Dataset

This dataset contains vector representations of faces extracted using the save_faces function. Each entry in the dataset consists of a vector representation of a face along with associated metadata.

Function Description

The save_faces function extracts vector representations of faces from images using the Face Recognition library. It takes as input a marked image (with detected faces), a list of face images (as NumPy arrays or similar), and corresponding names or identifiers for each face. The function iterates over the detected faces, computes their vector representations, and stores them along with the associated names in a dataset.

Dataset Contents

The dataset contains the following information for each entry:

• names: Names or identifiers corresponding to each face.
• face_vectors: Vector representations of the faces extracted using the save_faces function.

Usage Instructions

To use this dataset:

1. Obtain the marked_img, faces, and names inputs for the save_faces function.
2. Call the save_faces function with the provided inputs.
3. The function will generate a dataset containing the vector representations of faces along with their associated names.
4. The dataset will be saved to the Hugging Face Hub with the name "facenet-faces-vectors-dataset".

Example Usage

# Example usage of the save_faces function
marked_img, faces, names = ...  # Obtain the inputs
marked_img, faces = save_faces(marked_img, faces, names)