Submitted Abstracts
Abstract by Javid Pack
Personal Infomation
Presenter's Name
Javid Pack
Co-Presenters
None
Degree Level
Masters
Co-Authors
None
Abstract Infomation
Department
Computer Science
Faculty Advisor
Parris Egbert
Title
Real-time FLIP Fluid Simulation through Machine Learning Approximations
Abstract
3M - Fluids in computer generated imagery can add an impressive amount of realism to a scene. Fluid simulations are used extensively in modern computer animated films but are noticeably lacking in interactive applications such as computer games. The reason for the lack of fluid simulations in interactive applications is due to the prohibitively expensive computations required by the computer to simulate fluids using current algorithms. Recent research efforts have attempted to utilize machine learning techniques to approximate the movement of fluids in pursuit of real-time simulations. We propose integrating machine learning prediction techniques into a Fluid-Implicit-Particle (FLIP) simulation method and expect the hybrid nature of FLIP to facilitate further improvements in speed, visual quality, and flexibility when compared to current techniques.