afis 1.0.1

Classical AFIS library: minutiae extraction, matching, and quality (Python + optional native NBIS)

afis

Classical AFIS toolkit for fingerprint minutiae extraction, matching, and quality assessment. Pure Python by default; optional native NBIS backend (libafis_lfs) when built on macOS or Linux.

Install

make install

Uses uv. Equivalent:

uv sync --all-extras --group dev
make native

make native builds libafis_lfs.dylib (macOS) or libafis_lfs.so (Linux) into afis/_native/. Compiles in parallel (-j); install ccache for faster rebuilds.

Quickstart

from afis import FingerprintProcessor, match_templates

proc = FingerprintProcessor(backend="auto")
a = proc.process("finger_a.png")
b = proc.process("finger_b.png")
print(a.n_minutiae, a.header.image_quality)

result = match_templates(
    a.minutiae, b.minutiae,
    method="bozorth3",
    height_a=a.header.height,
    height_b=b.header.height,
)
print(result.score, result.decision)

CLI

uv run afis extract finger.png finger.xyt
uv run afis match a.xyt b.xyt --method bozorth3 --height 480
uv run afis quality finger.png

Develop

make test
make build

License

MIT. Copyright (c) 2026 Ashok Pant.