/TurtleFlow

Primary LanguageJupyter Notebook

TURTLEFLOW

Application to recognize and identify people based in a Convolutional Neural Network designed using TensorFlow.

To execute the application we have to follow the next steps

Create images

python src/take_face_square.py /<path>/<to>/<project>/data/shape_predictor_68_face_landmarks.dat

Load data to CSV

python src/create_convolutional_dataset.py /<path>/<to>/<project>/img

Execute Video Streaming and People Identification application.

jupyter notebook

Execute notebook called notebooks/Convolutional Neural Network.ipynb

Tech

TurtleFlow technologies requirements:

  • OpenCV - Open Source computer vision.
  • Dlib - Machine Learning library.
  • TensorFlow - TensorFlow is an Open Source Software Library for Machine Intelligence.
  • Pandas - Data manage python library.
  • Scikit-learn - Machine Learning library.
  • Numpy - Data manage library.

Doc

Here you can find a pdf document with the full research process and project architecture. Full Documentation

Video Tutorial

Video tutorial explaining the project structure. Video Tutorial

Demo

Real demo identifying two different people. Demo