Stochastic ray-tracing methods such as Monte-Carlo ray-tracing may be adapted to provide more efficient and more realistic ultrasound imaging systems and methods. Many random ray paths that are perturbed according to a probability distribution may be generated until they converge to the correct solution. New surface thickness and roughness models enable to reproduce complex ultrasound interactions such as multiple reflections and refractions. The contribution of each ray path may be further weighted to better simulate a beamformed ultrasound signal. Tracing many individual rays per transducer element is easily parallelizable on modern GPU computing architectures.