OpticalFlow

This repository consists of four different approaches for generating optical flow and can be demonstrated in Colab.

| Method | Paper | | ----------------- |:----------------------- | | FlowNet 2 | FlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networks | | Farneback | Two-Frame Motion Estimation Based on Polynomial Expansion | | Horn-Schunck | Determining Optical Flow | | Lucas-Kanade | An Iterative Image Registration Technique with an Application to Stereo Vision |

<p> The following optical flow was generated by FlowNet 2. </p> <div> <img src = 'src/flownet2_result.gif' width = '60%' height = '60%'> </div>

Usage

Use Colab's chrome extension or provide a ipynb URL at http://colab.research.google.com/github/ to open main.ipynb and walkthrough four different approaches.

Results on MPI-Sintel

<div align='center'> <img src = 'src/alley1_result.PNG' width = '90%' height = '90%'> </div>