The applications of machine learning methods on medical engineering have attracted lots of attention. In this research, we focus on piecewise linear technology in machine learning and its application on medical image/signal processing for noise, outliers, and large-scale data. Piecewise linear functions are the simplest extension to linear functions. On the one hand, piecewise linear functions can be analyzed and solved as a series of linear systems. On the other hand, a piecewise linear function is nonlinear on the whole and we can design a suitable one for desired properties. In this project, we will further analyze the properties, develop efficient algorithms, and then apply them in medical engineering. 

This project is supported by the Alexander von Humboldt Foundation.