The goal of this Computer Vision Research Project is to track and predict trajectory of Fast Moving Objects (FMOs). The main program is based on the RVOS code base, which is a temporal and spatial machine learning method used to segment videos. The original code can be found on the following repository: https://github.com/imatge-upc/rvos/tree/12016db67a18e4c4dea3f42f5b3c0b2623761c5b.
- The
rvos
subfolder is just the existing RVOS code we have cloned - it is there so we can synchronize the files that are needed in theRVOS_FMO.ipynb
Colab notebook. This is necessary because Google Colab is mounted on each of our own Google Drives, so the files used by the notebook have to be on each of our accounts if we want to run the notebook. - The
databases
subfolder is created by following the instructions inrvos/README.md
with the DAVIS dataset. - As such, the RVOS subfolder must be placed at
My Drive/FMO-Tracking
in each of our Google Drive workspaces.