The Cherenkov Telescope Array (CTA) will be the next-generation ground-based detector for gamma rays with very high energies. Telescopes will be located at one site each in both the northern and southern hemisphere. The arrays will comprise, in total, more than 100 telescopes of different sizes and designs. The sensitivity of CTA in its central energy range, i.e. between approximately 100 GeV and 1 TeV, will be driven by the performance of the Medium-Sized Telescopes (MSTs). This performance crucially depends on an exact alignment of the facets of the tessellated mirror surface of each telescope. In this contribution, an automated mirror alignment procedure for MSTs is presented. This procedure consists of two steps. First a rough mirror alignment is achieved with the socalled Bokeh method, which is based on the non-focused imaging of an artificial light source onto the Cherenkov camera plane. Afterwards, an optimal mirror alignment is achieved with an alignment procedure based on the focused imaging of stars. Here, the Bokeh method will be described in detail, including the hardware and software setups, devised technologies and pattern recognition with classical and neural network-based methods. Also results from star alignment procedures are given and compared to results from the Bokeh method. The performance of the presented approach is demonstrated with results obtained from measurements at the MST prototype installation in Berlin, Germany.