Optimizing Machine Learning to Increase Relevance in Photo Selection for Private School-Based Social Media

Machine learning, specifically deep learning, for face recognition applications has advanced significantly over the past 20 years [ref: science direct survey on deep based facial recognition]. There are many deep learning concepts pertinent to face image analysis and facial recognition, and there is active research in outstanding problems ranging from effective algorithms to handle variations in pose, age, illumination, expression, and heterogeneous face matching. And research continues in data sampling, training and modeling to better understand and address issues related to bias. In addition to the advances in research and application, there is recent, and overdue, greater awareness of the impact machine learning and AI have on broad aspects of our society and personal privacy. Within the dynamics of technological advances and societal value, applications such as Vidigami’s private and secure school-based social media platform are seeking to evaluate and implement new functionality and policies to maintain the trust and provide relevance to its user community from students, parents, teachers and school staff. In this proposal research project, the intern(s) will review and evaluate the latest applied techniques to improve facial recognition (e.g. including use of non-facial characteristics such as height and context) and provide greater relevance to school communities.

Faculty Supervisor:

Jiannan Wang

Student:

Sai Janaki Sathish

Partner:

Vidigami Media Inc.

Discipline:

Computer science

Sector:

Other

University:

Simon Fraser University