Fig. 12From: Applications of machine learning for simulations of red blood cells in microfluidic devicesPrediction error calculated as the distance of RBC centers from simulation B177f and centers predicted by a system of bases trained on B177e. The error values are large at first glance. This is due to the fact that the RBCs in channel B are moving more than 20 times faster than the cells in the channel A. If we reduce the errors 20-fold, they reach 50–60% of the error values in Fig. 9Back to article page