Skip to content

Latest commit

 

History

History
63 lines (53 loc) · 2.08 KB

README.md

File metadata and controls

63 lines (53 loc) · 2.08 KB
Raw

Live Facial Recognition

With DLIB and Face-Recognition.

At the time of development, November 16th, 2021, DLIB cannot be successfully installed with Conda or Pip, thus requiring a patch for the Jetson Nano 2GB.
You may skip these steps if dlib and face_recognition lib in Python3 work for you.

Otherwise, this tutorial on Medium helped me achieve a successful install.
I've extended the tutorial to create an isolated enviornment with Anaconda to prevent dependency hell.

Installing dependences

$ sudo apt-get update
$ sudo apt-get install python3-dev
$ sudo apt-get install build-dep python3
$ sudo apt-get install libblas-dev liblapack-dev libatlas-base-dev
$ sudo apt-get upgrade

Conda Enviornment

Note: I used Miniconda for venv

$ cd face_recog/
$ conda create --name face_recon --file envs/conda-env.txt
$ conda activate face_recon

Nvidia Jetson Patch for DLIB

Note the version number here...

$ git clone -b v19.21 https://github.com/davisking/dlib.git
$ cd dlib/
$ git am 0001-arch-for-Nano.patch
$ make

Patch on nvidia developer forums.

Compile and Install

Take note of your gcc version, mine is gcc version 7.5.0. Can be found via:

$ gcc -v

Enter the foler dlib and find setup.py file. This contains all the stuff describing how to install it.
Open it, and add the following line in it:

os.environ["CC"] = "gcc-7"

Building on Jetson

$ mkdir build && cd build
$ cmake --build . --config Release
$ sudo ldconfig

Installing Dependencies

For recog_known.py.
Important: Must in an active conda enviornment!

(face_recon) $ pip install -r envs/pip-req.txt
(face_recon) $ python3 recog_known.py