Skip to main content

Recognize faces from Python or from the command line

Project description

Face Recognition

| Recognize and manipulate faces from Python or from the command line
| the world's simplest face recognition library.

| Built using `dlib <>`__'s state-of-the-art face
| built with deep learning. The model has an accuracy of 99.38% on the
| `Labeled Faces in the Wild <>`__

| This also provides a simple ``face_recognition`` command line tool
that lets
| you do face recognition on a folder of images from the command line!

| |PyPI|
| |Build Status|
| |Documentation Status|


Find faces in pictures

Find all the faces that appear in a picture:


.. code:: python

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

Find and manipulate facial features in pictures

Get the locations and outlines of each person's eyes, nose, mouth and


.. code:: python

import face_recognition
image = face_recognition.load_image_file("your_file.jpg")
face_landmarks_list = face_recognition.face_landmarks(image)

| Finding facial features is super useful for lots of important stuff.
But you can also use for really stupid stuff
| like applying `digital
make-up <>`__
(think 'Meitu'):


Identify faces in pictures

Recognize who appears in each photo.


.. code:: python

import face_recognition
known_image = face_recognition.load_image_file("biden.jpg")
unknown_image = face_recognition.load_image_file("unknown.jpg")

biden_encoding = face_recognition.face_encodings(known_image)[0]
unknown_encoding = face_recognition.face_encodings(unknown_image)[0]

results = face_recognition.compare_faces([biden_encoding], unknown_encoding)

You can even use this library with other Python libraries to do
real-time face recognition:


See `this
example <>`__
for the code.



- Python 3.3+ or Python 2.7
- macOS or Linux (Windows not officially supported, but might work)

Installation Options:

Installing on Mac or Linux

First, make sure you have dlib already installed with Python bindings:

- `How to install dlib from source on macOS or
Ubuntu <>`__

Then, install this module from pypi using ``pip3`` (or ``pip2`` for
Python 2):

.. code:: bash

pip3 install face_recognition

| If you are having trouble with installation, you can also try out a
| `pre-configured
VM <>`__.

Installing on Raspberry Pi 2+

- `Raspberry Pi 2+ installation
instructions <>`__

Installing on Windows

While Windows isn't officially supported, helpful users have posted
instuctions on how to install this library:

- `@masoudr's Windows 10 installation guide (dlib +
face\_recognition) <>`__

Installing a pre-configured Virtual Machine image

- `Download the pre-configured VM
image <>`__
(for VMware Player or VirtualBox).


Command-Line Interface

| When you install ``face_recognition``, you get a simple command-line
| called ``face_recognition`` that you can use to recognize faces in a
| photograph or folder full for photographs.

| First, you need to provide a folder with one picture of each person
| already know. There should be one image file for each person with the
| files named according to who is in the picture:


Next, you need a second folder with the files you want to identify:


| Then in you simply run the command ``face_recognition``, passing in
| the folder of known people and the folder (or single image) with
| people and it tells you who is in each image:

.. code:: bash

$ face_recognition ./pictures_of_people_i_know/ ./unknown_pictures/

/unknown_pictures/unknown.jpg,Barack Obama

| There's one line in the output for each face. The data is
| with the filename and the name of the person found.

| An ``unknown_person`` is a face in the image that didn't match anyone
| your folder of known people.

Adjusting Tolerance / Sensitivity

| If you are getting multiple matches for the same person, it might be
| the people in your photos look very similar and a lower tolerance
| is needed to make face comparisons more strict.

| You can do that with the ``--tolerance`` parameter. The default
| value is 0.6 and lower numbers make face comparisons more strict:

.. code:: bash

$ face_recognition --tolerance 0.54 ./pictures_of_people_i_know/ ./unknown_pictures/

/unknown_pictures/unknown.jpg,Barack Obama

| If you want to see the face distance calculated for each match in
| to adjust the tolerance setting, you can use ``--show-distance true``:

.. code:: bash

$ face_recognition --show-distance true ./pictures_of_people_i_know/ ./unknown_pictures/

/unknown_pictures/unknown.jpg,Barack Obama,0.378542298956785

More Examples

| If you simply want to know the names of the people in each photograph
but don't
| care about file names, you could do this:

.. code:: bash

$ face_recognition ./pictures_of_people_i_know/ ./unknown_pictures/ | cut -d ',' -f2

Barack Obama

Speeding up Face Recognition

| Face recognition can be done in parallel if you have a computer with
| multiple CPU cores. For example if your system has 4 CPU cores, you
| process about 4 times as many images in the same amount of time by
| all your CPU cores in parallel.

If you are using Python 3.4 or newer, pass in a
``--cpus <number_of_cpu_cores_to_use>`` parameter:

.. code:: bash

$ face_recognition --cpus 4 ./pictures_of_people_i_know/ ./unknown_pictures/

You can also pass in ``--cpus -1`` to use all CPU cores in your system.

Python Module

| You can import the ``face_recognition`` module and then easily
| faces with just a couple of lines of code. It's super easy!

API Docs:
` <>`__.

Automatically find all the faces in an image

.. code:: python

import face_recognition

image = face_recognition.load_image_file("my_picture.jpg")
face_locations = face_recognition.face_locations(image)

# face_locations is now an array listing the co-ordinates of each face!

| See `this
example <>`__
| to try it out.

You can also opt-in to a somewhat more accurate deep-learning-based face
detection model.

| Note: GPU acceleration (via nvidia's CUDA library) is required for
| performance with this model. You'll also want to enable CUDA support
| when compliling ``dlib``.

.. code:: python

import face_recognition

image = face_recognition.load_image_file("my_picture.jpg")
face_locations = face_recognition.face_locations(image, model="cnn")

# face_locations is now an array listing the co-ordinates of each face!

| See `this
example <>`__
| to try it out.

| If you have a lot of images and a GPU, you can also
| `find faces in
batches <>`__.

Automatically locate the facial features of a person in an image

.. code:: python

import face_recognition

image = face_recognition.load_image_file("my_picture.jpg")
face_landmarks_list = face_recognition.face_landmarks(image)

# face_landmarks_list is now an array with the locations of each facial feature in each face.
# face_landmarks_list[0]['left_eye'] would be the location and outline of the first person's left eye.

| See `this
example <>`__
| to try it out.

Recognize faces in images and identify who they are

.. code:: python

import face_recognition

picture_of_me = face_recognition.load_image_file("me.jpg")
my_face_encoding = face_recognition.face_encodings(picture_of_me)[0]

# my_face_encoding now contains a universal 'encoding' of my facial features that can be compared to any other picture of a face!

unknown_picture = face_recognition.load_image_file("unknown.jpg")
unknown_face_encoding = face_recognition.face_encodings(unknown_picture)[0]

# Now we can see the two face encodings are of the same person with `compare_faces`!

results = face_recognition.compare_faces([my_face_encoding], unknown_face_encoding)

if results[0] == True:
print("It's a picture of me!")
print("It's not a picture of me!")

| See `this
example <>`__
| to try it out.

Python Code Examples

All the examples are available
`here <>`__.

Face Detection

- `Find faces in a
photograph <>`__
- `Find faces in a photograph (using deep
learning) <>`__
- `Find faces in batches of images w/ GPU (using deep
learning) <>`__

Facial Features

- `Identify specific facial features in a
photograph <>`__
- `Apply (horribly ugly) digital
make-up <>`__

Facial Recognition

- `Find and recognize unknown faces in a photograph based on
photographs of known
people <>`__
- `Compare faces by numeric face distance instead of only True/False
matches <>`__
- `Recognize faces in live video using your webcam - Simple / Slower
Version (Requires OpenCV to be
installed) <>`__
- `Recognize faces in live video using your webcam - Faster Version
(Requires OpenCV to be
installed) <>`__
- `Recognize faces in a video file and write out new video file
(Requires OpenCV to be
installed) <>`__
- `Recognize faces on a Raspberry Pi w/
camera <>`__
- `Run a web service to recognize faces via HTTP (Requires Flask to be
installed) <>`__

How Face Recognition Works

| If you want to learn how face location and recognition work instead of
| depending on a black box library, `read my
article <>`__.


- The face recognition model is trained on adults and does not work
very well on children. It tends to mix
up children quite easy using the default comparison threshold of 0.6.

Deployment to Cloud Hosts (Heroku, AWS, etc)

| Since ``face_recognition`` depends on ``dlib`` which is written in
C++, it can be tricky to deploy an app
| using it to a cloud hosting provider like Heroku or AWS.

| To make things easier, there's an example Dockerfile in this repo that
shows how to run an app built with
| ``face_recognition`` in a `Docker <>`__
container. With that, you should be able to deploy
| to any service that supports Docker images.

Common Issues

Issue: ``Illegal instruction (core dumped)`` when using
face\_recognition or running examples.

| Solution: ``dlib`` is compiled with SSE4 or AVX support, but your CPU
is too old and doesn't support that.
| You'll need to recompile ``dlib`` after `making the code change
here <>`__.

``RuntimeError: Unsupported image type, must be 8bit gray or RGB image.``
when running the webcam examples.

Solution: Your webcam probably isn't set up correctly with OpenCV. `Look
here for
more <>`__.

Issue: ``MemoryError`` when running ``pip2 install face_recognition``

| Solution: The face\_recognition\_models file is too big for your
available pip cache memory. Instead,
| try ``pip2 --no-cache-dir install face_recognition`` to avoid the

``AttributeError: 'module' object has no attribute 'face_recognition_model_v1'``

Solution: The version of ``dlib`` you have installed is too old. You
need version 19.7 or newer. Upgrade ``dlib``.

``Attribute Error: 'Module' object has no attribute 'cnn_face_detection_model_v1'``

Solution: The version of ``dlib`` you have installed is too old. You
need version 19.7 or newer. Upgrade ``dlib``.

Issue: ``TypeError: imread() got an unexpected keyword argument 'mode'``

Solution: The version of ``scipy`` you have installed is too old. You
need version 0.17 or newer. Upgrade ``scipy``.


- Many, many thanks to `Davis King <>`__
(`@nulhom <>`__)
for creating dlib and for providing the trained facial feature
detection and face encoding models
used in this library. For more information on the ResNet that powers
the face encodings, check out
his `blog
post <>`__.
- Thanks to everyone who works on all the awesome Python data science
libraries like numpy, scipy, scikit-image,
pillow, etc, etc that makes this kind of stuff so easy and fun in
- Thanks to `Cookiecutter <>`__
and the
`audreyr/cookiecutter-pypackage <>`__
project template
for making Python project packaging way more tolerable.

.. |PyPI| image::
.. |Build Status| image::
.. |Documentation Status| image::
.. |image3| image::
.. |image4| image::
.. |image5| image::
.. |image6| image::
.. |image7| image::
.. |known| image::
.. |unknown| image::


1.2.0 (2018-02-01)

* Fixed a bug where batch size parameter didn't work correctly when doing batch face detections on GPU.
* Updated OpenCV examples to do proper BGR -> RGB conversion
* Updated webcam examples to avoid common mistakes and reduce support questions
* Added a KNN classification example
* Added an example of automatically blurring faces in images or videos
* Updated Dockerfile example to use dlib v19.9 which removes the boost dependency.

1.1.0 (2017-09-23)

* Will use dlib's 5-point face pose estimator when possible for speed (instead of 68-point face pose esimator)
* dlib v19.7 is now the minimum required version
* face_recognition_models v0.3.0 is now the minimum required version

1.0.0 (2017-08-29)

* Added support for dlib's CNN face detection model via model="cnn" parameter on face detecion call
* Added support for GPU batched face detections using dlib's CNN face detector model
* Added to examples
* Added to examples
* Added to examples
* dlib v19.5 is now the minimum required version
* face_recognition_models v0.2.0 is now the minimum required version

0.2.2 (2017-07-07)

* Added --show-distance to cli
* Fixed a bug where --tolerance was ignored in cli if testing a single image
* Added to examples

0.2.1 (2017-07-03)

* Added --tolerance to cli

0.2.0 (2017-06-03)

* The CLI can now take advantage of multiple CPUs. Just pass in the -cpus X parameter where X is the number of CPUs to use.
* Added example
* Improved CLI tests to actually test the CLI functionality
* Updated to capture in rgb (not bgr) format.

0.1.14 (2017-04-22)

* Fixed a ValueError crash when using the CLI on Python 2.7

0.1.13 (2017-04-20)

* Raspberry Pi support.

0.1.12 (2017-04-13)

* Fixed: Face landmarks wasn't returning all chin points.

0.1.11 (2017-03-30)

* Fixed a minor bug in the command-line interface.

0.1.10 (2017-03-21)

* Minor pref improvements with face comparisons.
* Test updates.

0.1.9 (2017-03-16)

* Fix minimum scipy version required.

0.1.8 (2017-03-16)

* Fix missing Pillow dependency.

0.1.7 (2017-03-13)

* First working release.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

face_recognition-1.1.0.tar.gz (3.2 MB view hashes)

Uploaded Source

Built Distribution

face_recognition-1.1.0-py2.py3-none-any.whl (19.7 kB view hashes)

Uploaded Python 2 Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page