Describing complex phenomena directly is difficult to do. In those situations, an alternative, implicit definition can often help us model the phenomenon and perform simulations of it.
Ordinary Differential Equations (ODEs) do this by describing how a system changes, that is, they present equations that capture the "dynamics" of the system.
In this tutorial, we will begin to explore what learning these dynamics through neural network training allows us to do. Examples include making density estimation through normalizing flows continuous and allowing time-series models to take in data sampled at irregular intervals.
The majority of deaf-and-mute people use sign language produced by body actions such as hand gestures, body motion, eyes and facial expressions to communicate amongst each other and with non-impaired people in their daily life. However, it has become a barrier for mute and deaf communities which intend to integrate into society. To bridge the communication gap, a hand gesture recognition system for Sign Language Recognition (SLR) is required.
This project aims to design a real-time vision-based hand gesture recognition system with machine learning techniques, which potentially makes deaf-and-mute people life easier. In practice, signs are always continuously spelt words mixing both dynamic and static gestures, so the wanted recognition system should be able to recognize both dynamic and static gestures in ASL with promising accuracy.