% Code and data for ICCV 2009 poster: Seeing through water: image restoration using model-based tracking 
% Created by Yuandong Tian, Oct. 5, 2009
% Email: yuandong@andrew.cmu.edu

% figure_*.m            -> matlab scripts to make the corresponding figures in the paper. 
% init.m                -> scripts that generates the synthesis dataset, including distorted frames and water heights. random seed included.
% waterTracking.m       -> The water tracking procedure. 
% warp.mat              -> The water bases. 
% expdata_*.mat         -> real experiment data, each is 61-frames long. 
% exp_computed_*.mat    -> computed coefficients on the experiments. All coefficients are produced by waterTracking.m
% 