RealTime Video Face Detection

Demo Link

Installation

• Install Homebrew

  • /usr/bin/ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"
    

  • vim ~/.bash_profile
    export PATH=/usr/local/bin:$PATH
    

    source ~/.bash_profile
    

• Install python

  • brew install python python3
    cd /usr/local/bin
    ln -s ../Cellar/python/2.7.14/bin/python2 python
    

• Install ffmpeg

  • brew install ffmpeg \
      --with-tools \
      --with-fdk-aac \
      --with-freetype \
     --with-fontconfig \
     --with-libass \
     --with-libvorbis \
     --with-libvpx \
     --with-opus \
     --with-x265

• Install opencv

  • brew tap homebrew/science
    brew install opencv3 --with-contrib --with-python3
    

• Install boto3, watchdog

  •   pip2 install boto3 watchdog
    

  or python -m pip install boto3 watchdog

Execution - Recognizing Faces in a Streaming Video

Clone this repo:

git clone [email protected]:imyoungyang/video-streaming.git

Step1: IAM Role & SNS Topic

Create an IAM service role to give Rekognition Video access to your Kinesis video streams and your Kinesis data streams. SNS Topic to recieve the reginition name.

python iam-role-helper.py --create
python sns-helper.py --create

Step2: Create Collection

python collection-helper.py --create

Step3: Add faces to a collection

python index_faces.py ./young-yang.jpg Young

Step4: Create a Kinesis Video Stream

python video-stream-helper.py --create

Step5: Create a Kinesis Data Stream

python data-stream-helper.py --create

Step6: Create the stream processor

• run command to create a stream processor.

  python rekognition-process.py --create

• run command to check process status

  aws rekognition describe-stream-processor --name appStreamProcessor-videoFaceRek

Step7: Start the stream processor

• run command python rekognition-process.py --start to start the process

Step8: Start video stream

• Open terminal and exeucte the upload to kinesis videos

  • python watch_for_changes.py

• Execute face detection in another terminal

  • python face-detection-multi-files.py

Step9: Consume the analysis result

• run command python get-rekognition-result.py

Reference

• AWS developer guide: recognize faces in a video stream