The goal of this project is to detect and reveal temporal changes in videos that are nearly impossible with the naked eye. The method used for this is known as Eulerian Video Magnification. It takes in a standard video captured by a normal video camera as input, it then applies spatial decomposition and temporal filtering to individual frames. The resulting output is then amplified or magnified which reveals minute changes which are otherwise impossible to see with naked eyes. This method enables us to visualise how human’s face changes its colour when blood flows through it and also how detect minute movements by amplifying the input. This method is useful in various fields – video editing, video analysis for medical purposes, sports, products quality control and many more
The code and other resources can be found when you view this project on GitHub. Try running the given code on the two input videos or use your own video.