face_recognition Recognition

Face-recognition is an open-source Python library that provides face detection, face alignment, and face recognition capabilities. The official documentation can be found here

Installation

The face_recognition requires the following pre-requisites #. dlib

dlib Installation

To install dlib, you need to ensure that you meet the following specifications:

Operating System: dlib is compatible with Windows, macOS, and Linux operating systems.

Python Version: dlib works with Python 2.7 or Python 3.x versions.

Compiler: You need a C++ compiler to build and install dlib. For Windows, you can use Microsoft Visual C++ (MSVC) or MinGW. On macOS, Xcode Command Line Tools are required. On Linux, the GNU C++ Compiler (g++) is typically used.

Dependencies: dlib relies on a few external dependencies, including Boost and CMake. These dependencies need to be installed beforehand to successfully build dlib.

Windows

The official installation instructions are found here

  • To install the library, first you need to install the dlib library

Installation instructions are here

  1. Download CMake windows installer from here

  2. While installing CMake select Add CMake to the system PATH to avoid any error in the next steps.

  3. Install Visual C++ , if not installed previously

Then run the following commands to install the face_recognition #. cmake installation .. code-block:

pip install cmake

  1. dlib installation

pip install dlib

  1. face_recognition

mac OS installation

mac installation is pretty straightforward.

pip install dronebuddylib[FACE_RECOGNITION]

Usage

In order to proceed with the face recognition , the algorithm needs encoding s of the known faces. The library has a method that is specifically designed to add these faces to the memory.

engine_configs = EngineConfigurations({})
image = cv2.imread('test_clear.jpg')
engine = FaceRecognitionEngine(FaceRecognitionAlgorithm.FACE_RECC, engine_configs)
result = engine.remember_face(image, "Jane")

You can check if the images and names are added to the library by simply going to the location where the library is installed.

  1. SPEECH_RECOGNITION_VOSK_LANGUAGE_MODEL_PATH - This is the path to the model that you have downloaded. This is a compulsory parameter if you are using any other language. If this is not provided, the default model will be used. The default model is the english model ( vosk-model-small-en-us-0.15 ). Vosk supported languages can be found here.

Output

The output will be a list of names, if no people are spotted in the frame empty list will be returned. If people are spotted but not recognized, ‘unknown’ will be added as a list item.

Resources