This project processes video to amplify and segment motion.

1. amplify motion using a phase-based eularian approach using complex steerable pyramids
2. create a velocity vector field from the amplified motion
3. use the velocity vector field, along with other feature of the video to segment the video into meaningful parts.

data - where the original video, and all intermediate data is stored.

experiments - contains scripts performing step by step analysis

analysis pipeline : motionAmp - > mVectorField -> segment

tests - contains functions for generating synthetic test input for validation

utils - functions that are used in multiple places or are helper functions

​