---
id: face-recognition
version: "1.3.0"
license: MIT license
license_treatment: permissive
maintenance: active
---

# face-recognition — Recognize faces from Python or from the command line

License: permissive · Maintenance: active · Downloads: 209.1K/mo

## What it is and what it does

face-recognition wraps dlib's deep-learning face recognition model to provide a simple Python interface for detecting faces, extracting facial landmarks, and comparing face encodings to identify individuals. It includes both a Python API for programmatic use and a command-line tool for batch processing folders of images.

The library handles the core tasks of face detection (finding where faces appear in an image), facial feature location (eyes, nose, mouth, chin), and face identification (comparing unknown faces against known encodings to determine matches). It supports parallel processing across multiple CPU cores and allows tolerance tuning to adjust match sensitivity.

Use it for:

- Batch identify people in a folder of photos by comparing against a reference set of known individuals.
- Extract and store face encodings from a database of known people for later comparison and matching.
- Detect all faces in an image and extract their landmark coordinates for facial feature analysis or digital effects.
- Build a real-time face recognition system by processing video frames with parallel CPU processing.
- Adjust recognition sensitivity by tuning tolerance values when similar-looking people cause false matches.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Detects, locates, and identifies faces in images using deep learning, with both Python API and command-line interface for batch processing.

Yes, if you are on macOS or Linux and can install dlib. The package is actively maintained, has no known vulnerabilities, uses a permissive MIT license, and provides a straightforward API backed by a well-regarded deep learning model. Install friction is moderate due to dlib's compilation requirement, but the library itself installs cleanly. Not suitable for Windows without unofficial workarounds.

## Install

pip install face-recognition
uv add face-recognition
poetry add face-recognition

## Installing face-recognition

Before you install: Low friction installation on macOS and Linux; depends on dlib, which requires compilation from source. Windows is not officially supported. The package is actively maintained with recent commits, though the latest release is from 2020.

License in practice: MIT license permits commercial and private use, modification, and distribution with minimal restrictions—suitable for most projects.

Quickstart:

pip install face-recognition

import face_recognition
image = face_recognition.load_image_file("photo.jpg")
face_locations = face_recognition.face_locations(image)

dlib must be installed with Python bindings before installing face-recognition; requires macOS or Linux (Windows not officially supported).

Verify before relying:

- Whether the 99.38% accuracy figure on Labeled Faces in the Wild benchmark remains current for version 1.3.0.
- Real-world performance and accuracy on diverse face datasets outside the benchmark.
- GPU acceleration requirements and CUDA support status for deep-learning face detection model.

## Package facts

- License: MIT license (permissive)
- Python support: unspecified
- Install friction: low
- Maintenance: active
- Downloads: 209.1K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

face detection and recognition, identify people in photos, facial feature detection, face comparison and matching, batch face recognition, face encoding and distance, facial landmark detection, face-detection, deep-learning, computer-vision

[View on SkillFed](https://skillfed.io/packages/face-recognition) · [View on PyPI](https://pypi.org/project/face-recognition/)