Machine learning based privacy mechanisms

• Main Author: Lim, Royce Jin Feng
• Other Authors: Tay, Wee Peng
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2020
• Subjects: Engineering::Electrical and electronic engineering::Computer hardware, software and systems
• Online Access: https://hdl.handle.net/10356/140365
• Institution: Nanyang Technological University

The objective of this project is to learn latent representations using a Machine Learning approach for image sanitization in Smart Homes surveillance cameras. Three Autoencoder Machine Learning models are explored in this project 1) Adversarial Autoencoder 2) Variational Autoencoder 3) Variational Fair Autoencoder. These autoencoders are able to learn latent representations of the input data, which can be processed to encourage separation between the input data and private information which are classified in the model as sensitive variables. This provides the capability of sanitizing the image of the Smart Home user’s private information.