This is demo of the Blur Eqalisation Technique (BET) presented in XianSubbarao2006. It uses the breakfast dataset from WatanabeNayar1998 and POV-Ray generated datasets, patio and office, presented in DevernayPujadesVijay2012. The last two datasets feature exact ground truths.

Download the repository and use IPython to run:

• demo_depth_bet.py, it generates a synthetic dataset;
• breakfast_data_bet.py, it uses the breakfast dataset;
• povr_data_bet.py, it can use the patio or office dataset and it evaluates the method using respective the ground truth.