black and white images

[benjaminabruzzo]

Is there a config that needs to be changed for black and white images?

[tomas-gajarsky]

No special configuration is needed! You can use black and white images exactly like color images:

Initialize analyzer with default config

cfg = facetorch.config.get_cfg_path()
analyzer = facetorch.FaceAnalyzer(cfg)

Works with grayscale images - no special config needed

response = analyzer.run(
    path_image="path/to/your/grayscale_image.jpg",
    batch_size=1,
    fix_img_size=True,
    return_img_data=False,
    include_tensors=True
)

The UniversalReader supports grayscale images from all input sources:

File paths (.jpg, .png, etc.)
URLs
PIL Image objects in "L" mode
Bytes data containing grayscale images
NumPy arrays (though these need to be 3-channel RGB format)
Summary
No configuration changes are required for black and white images in facetorch. The library automatically handles the conversion from grayscale to RGB internally, ensuring compatibility with all the face detection and analysis models that expect RGB input.