face-feature-parser 1.0.1

A Python library for face parsing and feature extraction

FaceParser

A Python library for face parsing and feature extraction using BiSeNet. This library provides tools for facial feature detection, segmentation, and analysis.

Features

• Facial feature parsing and segmentation
• Feature coordinate extraction
• Facial metrics calculation
• CLI tool for easy model management and image processing
• Visualization tools for parsed features

Installation

pip install faceparser

After installation, you need to download the model weights:

faceparser install

Usage

Command Line Interface

1. Install model:

faceparser install

2. Check model installation:

faceparser check

3. Parse an image:

faceparser parse path/to/image.jpg -o output_directory --save-masks

Python API

from faceparser import get_coordinates

# Parse an image
coordinates = get_coordinates("path/to/image.jpg")

# Get feature embeddings
embeddings = coordinates.get_all_embeddings()

# Get face metrics
metrics = coordinates.get_relative_metrics()
symmetry = coordinates.get_symmetry_score()

# Access individual feature metrics
for feature, area in metrics['feature_areas'].items():
    print(f"{feature} area: {area:.3f}")

Model Details

The face parsing model is based on BiSeNet architecture and trained on the CelebAMask-HQ dataset. It can detect and segment the following facial features:

• Skin
• Nose
• Eyes (left and right)
• Eyebrows (left and right)
• Ears (left and right)
• Mouth
• Upper and lower lips
• Hair
• Hat
• Eyeglasses
• Earrings
• Necklace
• Neck
• Clothes

Feature Metrics

The library provides various metrics for facial analysis:

• Feature embeddings (normalized position, size, area, aspect ratio)
• Face symmetry score
• Feature areas (normalized)
• Normalized distances between features
• Relative positions
• Feature aspect ratios

License

This project is licensed under the MIT License - see the LICENSE file for details.#\x00 \x00f\x00a\x00c\x00e\x00p\x00a\x00r\x00s\x00e\x00r\x00 \x00 \x00