Compressed sensing-based smoke simulation

As we known, high-dimension smoke simulation is time consuming because of large amount of data sampling and computing. We present a method that introduces compressed sensing to smoke simulation. This method could significantly reduce the sampling data and the simulation time, while giving a convincing effect. We use the wavelet basis as compressing basis, and OMP as the reconstruction method, and reconstruct high-dimension smoke from low-dimension data frame by frame. Our method is fast and easy to implement, and the data sampling could not obey the Nyquist sampling theorem which could reduce the amount of data.