VOSEvaluator

An Ipython-widgets based monitoring tool for Semi-Supervised Video Object Segmentation models

VOSEvaluator.mov

The evaluation code is based on DAVIS 2017 Semi-supervised and Unsupervised evaluation package

Usage:

In your conda environemt, run monitoring_vos_model.ipynb

Specify paths to your data in:

frame_path = 'Path/To/JPEGImages' #Replace with frames dir
gt_path = 'Path/To/Annotations' #Replace with annotations dir
results_path = 'Path/To/Results' #Replace with results dir
saved_evals = 'Path/To/SavedEvaluation' #Replace with saved_evaluation dir

And run the Cell