MSP Lab researches theories and applications for signal processing, whose keywords are "Multidimensional Multimedia Signal Processing". In this page, the representative research themes are indicated as follows:
Graph Signal Processing (GSP)
GSP is one of the emerging topics in signal processing, and its research publications are grown rapidly. The difference between GSP and regular signal processing is that the structure of signals is explicitly considered in GSP. The characteristic enables GSP to efficiently process a multidimensional signal in comparison of regular signal processing. However, GSP is not substantially studied yet so that MSP Lab develops its theories. Moreover, we apply the developed theories to some applications.
Some image processing applications have been presented, such as denoising, inpainting, background modeling, colorization, enhancement, stylization, pencil drawing and so on. In general, the energy of an image is localized to low frequency components in its frequency domain, and the characteristic is used for solving some image processing problems. By using signal processing perspective, MSP Lab introduces new theories for image processing and efficiently resolves some image processing problems.