DeepFace: a lightweight Python library that implements facial age, gender, emotion, race recognition

General Introduction

DeepFace is a lightweight Python library for facial recognition and facial attribute analysis (including age, gender, emotion, and ethnicity). It integrates several state-of-the-art facial recognition models such as VGG-Face, FaceNet, OpenFace, DeepFace, DeepID, ArcFace, Dlib, SFace, and GhostFaceNet.DeepFace not only enables high-precision facial recognition, but also performs detailed analysis of facial attributes. The library is designed with the goal of simplifying the process of facial recognition by enabling developers to easily invoke its features for facial verification, lookup and analysis.

Function List

• Facial Recognition: Highly accurate facial recognition through multiple models.
• Facial Attribute Analysis: analyzes the face for age, gender, mood, and ethnicity.
• Facial Verification: Verifies that two facial images belong to the same person.
• Find function: Finds faces in the database that match the input image.
• Supports multiple input formats: supports image paths, numpy arrays and base64 encoded images.
• Efficient facial embedding storage: use pickle files to store facial embeddings for faster lookups.
• Flexible installation: support for installation via PyPI and source code.

Using Help

Installation process

The DeepFace library can be installed in two ways:

1. Installation via PyPI:

pip install deepface

2. Installation via source code:

git clone https://github.com/serengil/deepface.git
cd deepface
pip install -e .

usage example

Once installed, you can import and use the features of the DeepFace library with the following code:

from deepface import DeepFace

facial verification

Verify that the two facial images belong to the same person:

result = DeepFace.verify(img1_path="img1.jpg", img2_path="img2.jpg")
print(result["verified"])

facial recognition

Find faces in the database that match the input image:

result = DeepFace.find(img_path="img.jpg", db_path="database/")
print(result)

Facial Attribute Analysis

Analyze faces for age, gender, mood, and ethnicity:

result = DeepFace.analyze(img_path="img.jpg", actions=['age', 'gender', 'emotion', 'race'])
print(result)

Advanced Features

The DeepFace library also provides some advanced features such as batch prediction, face extraction options, and more. Below are some examples of how to use some of the advanced features:

Batch prediction

results = DeepFace.analyze(img_paths=["img1.jpg", "img2.jpg"], actions=['age', 'gender', 'emotion', 'race'])
print(results)

Face Extraction Options

faces = DeepFace.extract_faces(img_path="img.jpg", target_size=(224, 224), grayscale=False)
print(faces)

common problems

1. How can I increase the speed of recognition?
   • Finding can be accelerated by pre-computing and storing facial embeddings.
   • Use efficient hardware (e.g., GPUs) for computation.
2. How to deal with low resolution images?
   • It is possible to useresampleparameter to improve the quality of low-resolution images.
3. How to handle multiple faces in a large image?
   • utilizationmax_facesparameter limits the number of faces processed.

With the above detailed usage help, users can easily get started with the DeepFace library for facial recognition and attribute analysis.