Facial Analysis: Detector 2D (Python API)

• Introduction
• Trained Models
• Requirements
• Status
• Related projects

Introduction

facialAnalysisDetection2D module use deepFace python API. The module analyze faces using pre-trained models and adds facial analysis doing prediction about some features like gender, age, race and emotions. Also use YARP to send video source pre and post-procesed. Also admits YARP source video like input. This module also publish detection results in YARP port. Documentation available on docs

Trained Models

facialAnalysisDetection2D requires images source to detect. First run program will download pre-trained models about features to detect:

1. Execute programs/facialAnalysisDetection2D.py, to start the program.

python3 facialAnalysisDetection2D.py

3. Connect video source to facialAnalysisDetection2D.

yarp connect /videoSource /facialAnalysisDetection2D/img:i

NOTE:

• Video results are published on /facialAnalysisDetection2D/img:o
• Data results are published on /facialAnalysisDetection2D/data:o

Requirements

facialRecognitionDetection2D requires:

• Install OpenCV 3.0.0+
• Install YARP 2.3.XX+
• Install pip
• Install deepFace:

pip3 install deepface

NOTE: This module use Tensorflow backend.

Possible errors: deepFace requires Python 3.5.5 but Ubuntu 16.04 use Python 3.5.2 by default. You can add Python 3.6 with PPA repository like:

sudo add-apt-repository ppa:deadsnakes/ppa
sudo apt-get update
sudo apt-get install python3.6

To install pip in Python 3.6:

curl https://bootstrap.pypa.io/get-pip.py | sudo python3.6

You can install the package with:

python3.6 -m pip install deepface

If pip change the link to Python 3.6 by default, you can revert to link pip to Python 3.5 with:

python3 -m pip install --upgrade --force pip

Tested on: ubuntu 14.04, ubuntu 16.04, ubuntu 18.04, lubuntu 18.04 and raspbian.

Status

Related projects

• serengil: deepFace project