Description

This repository is prepared for Grace Hopper Celebration workshop session: Confronting Algorithmic Bias in Artificial Intelligence. The repository contains implementation of mobilenet-ssd for face detection written in keras. The model training portion is borrowed from https://github.com/bruceyang2012/Face-detection-with-mobilenet-ssd

Required Software

Mac OS

Prerequisites

macOS 10.12.6 Sierra or above (Click Apple log in the top left corner and select 'About This Mac' to get the OS version)

Instructions

1. Download and install Python3 from official site. If you already have existing Python3, please make sure the version is 3.6.5 by running 'python3 --version'

2. Upgrade pip to the latest

python3 -m pip install --user --upgrade pip

3. Install virtualenv

python3 -m pip install --user virtualenv

4. Create a virtualenv

python3 -m virtualenv env

5. Activating the virtualenv

source env/bin/activate

6. Install required python packages

pip3 install keras tensorflow opencv-python jupyter imutils sklearn pillow beautifulsoup4 matplotlib tqdm

Troubleshooting steps

1. If you see the following error, make sure the OS version is 10.12.6 or above. To check the version, click Apple log in the top left corner and select 'About This Mac'

ImportError: dlopen(/usr/local/lib/python3.6/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so, 6): Symbol not found: _SecKeyCopyExternalRepresentation Referenced from: /usr/local/lib/python3.6/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so (which was built for Mac OS X 10.12)

2. If you do not see the live video window, wait for up to 10 seconds depending on the processing power of the laptop. It is possible that the window may be hidden.

3. If you run into issues while running Jupyter notebook, please kill the Jupyter notebook and restart the virtualenv enviornment

Ubuntu

Prerequisites

Ubuntu 16.04 or above

1. Install Python 3

sudo apt update && sudo apt install python3.6

2. Install required python packages

pip3 install keras tensorflow opencv-python jupyter imutils sklearn pillow beautifulsoup4 matplotlib tqdm

Live Inference

The repository comes with a pre-trained model to detect faces from the webcam video. Follow the steps to run this inference

1. Download or clone this repository by clicking the 'clone or download' button on this page.
2. Unzip the downloaded zip file.
3. Open a terminal and run the following command:

cd <download directory>

jupyter notebook

4. Jupyter will open a new tab or window on your default browser. Click on 'live.ipynb', and click 'Run' on each cell to execute. Once you reach the last cell, wait a few seconds for a new window to open up. You can see the face detection in action!

5. Press 'p' key on the video window to pause and 'Esc' key to stop the inference.

Challenges

1. Count the number of people in the room.
2. Capture any new face and save it as an image on local disk.
3. Sound an alarm when an intruder is detected!

Training Model

This is outside the scope of the workshop and is included for the benefit of the audience and for the fullness of the workshop. Please refer to the original GitHub repository for more information on the training: https://github.com/bruceyang2012/Face-detection-with-mobilenet-ssd

License

MIT LICENSE