Researcher at the Centre for Mobile Innovation at Sheridan College

Early research work in 2019 where I looked into fall detection in elderly care facilities and doing emotion recognition with TensorFlow 2. Learnt about machine learning and how to effectively work as a team of 1.

Alec Di Vito
Alec Di Vito 1 min read
Researcher at the Centre for Mobile Innovation at Sheridan College

During the year of 2019 I worked as a Researcher and a Research Assistant learning about machine learning. During my time there I independently worked to build a fall detection model and an facial expression classifier.

The first project was predicting when a fall occurred while in a home. The idea was to use google human pose detection library and a machine learning support vector machine (SVM) to predict when the fall occurred. Through my implementation and testing I was able to get an the accuracy of around 80%. The setup required depended a lot on where the camera was positioned and the points of the human body were used in the SVM to then handle the prediction.

The second project was to build a real time facial expression classifier for marketing purposes. I built a machine learning model that reached a 75% accuracy in the evaluation dataset. I built a neural network which would feed in human faces and predict their emotion depending on the emotions they showed.

The defining learning I had during this experience was how working alone feels. During my year as a Researcher, I was left alone to learn and produce results at the same time. I found it surprising tough and released how much help a team of people actual help to be able to deliver a solution.