Deep Learning

In recent years, deep learning has been spotlighted as a powerful tool to help overcome long-lasting problems in various fields including fluid mechanics.

Various flow visualization techniques based on digital image processing, such as shadowgraph and PIV, have been actively used and are working well in the experimental study of fluid dynamics, but there is still room for improvement in terms of time and human cost.

Currently, in the purpose of increasing robustness of image processing techniques in variety of experimental environments and reducing the overall processing cost, deep learning models are applied for the various visualization techniques.