Our company plans to develop vehicle-mounted security and active safety terminal to realize national mandatory DMS, lane shift, forward collision function and other industrial applications.In the implementation of these image algorithms, some algorithms need to use deep learning, while others use traditional image algorithms.Is there a solution recommendation?