Facial Recognition with IBM Watson

Project Description:

The IBM Watson suite is a Unity plugin that allows you to train classifiers and use learning algorithms located on the IBM cloud servers. Using these algorithms you can create programs that can recognize faces, engage in conversations, and process languages and information. My project seeks to configure a program, using IBM’s Visual Recognition service, to identify and classify my face and be able to detect it from others’ faces. This program would be used as the login information for an app or game and should be portable enough to assign to any application. The IBM Watson service is free to use, as is the plugin and Unity.

Methods:

I will be using Unity as the environment in which to contain the facial recognition software and display the webcam and related components. Despite using existing algorithms, it’s up to me to figure out how to program the facial recognition and train the classifiers. The programming is done in Unity C# using a related IDE like Visual Studio Code. I will be using the IBM Watson website to learn how to develop the program, as well as various online tutorials and support.

Project Deliverables:

  • Unity environment, assets and C# code
  • Documentation on IBM Watson, Visual Recognition service, code used
  • Relevant flowcharts and diagrams

Schedule/Calendar:

Work Breakdown Structure
1.1 Research 2.1 Setup 3.1 Design 4.1 Programming
1.1.1 Preliminary Research 2.1.1 Setup TensorFlow Environment 3.1.1 Design Maze for Unity 4.1.1 Program Interactions
1.1.2 Preliminary Modeling & Programming 2.1.2 Setup Python Notebook 3.1.2 Design NPC/Enemy and Misc. Models 4.1.2 Program A.I. Behavior
  2.1.3 Setup ML-Agents Plugin 3.1.3 Configure Model Types and Collision Detection 4.1.3 Decide and Test A.I. Models
  2.1.4 Test Setup and Troubleshoot Compatibility 3.1.4 Testing and Troubleshooting 4.1.4 Run Iterations of Models