Forschungsthemen am Lehrstuhl für Medientechnik

Der Lehrstuhl für Medientechnik (LMT) beschäftigt sich in Forschung und Lehre mit folgenden Themen:

3D Image Analysis and Synthesis Mobile Multimedia Networked Multimedia Media Compression

A major focus of our research activities is in the area of compression and coding of multimedia/multimodal (audio, video, haptic) information. The main goals of our work are to improve the rate-distortion performance and error resiliency of today's compression schemes. Recently, we have intensified our activities in the field of data reduction and compression for haptic data streams. This is a particularly challenging research topic due to the extremely low delay and stability requirements of haptic data communication. Recent topics are:

  • Rate-distortion optimized image and video compression
  • Compression of color filter array (Bayer-pattern) video sequences for single chip video cameras
  • Error resilient video communication using proxies
  • Error resilient short-range point-to-point video communication with very low delay
  • Distributed video coding / Wyner-Ziv video coding
  • Joint source-channel coding
  • Compression of haptic data for telepresence and teleaction
  • Compression of haptic data for haptic recording
  • Psychophysics-based data reduction for haptic data streams

Research assistants in this area:

A major focus of our research activities is in the area of networked multimedia systems. Our goal is to develop systems that are easy to use and that allow the user to access the desired multimedia content at any time from any place with any kind of device. To this end we study adaptive coding, adaptive transmission and adaptive display of multimedia information. A major expertise in our group is to design algorithms and systems for networked multimedia such that the user satisfaction is maximized. Recent topics are:

  • Internet media streaming
  • Packet path diversity and adaptive media playout
  • Rate-distortion optimized packet scheduling and rate shaping
  • Media distribution in content delivery networks
  • Media distribution in peer-to-peer networks
  • Multimodal data communication for telepresence and teleaction
  • Internet Protocol TV (IPTV)
  • Video-on-demand using early playout scheduling
  • Proxy-support for networked multimedia systems
  • In-car multimedia communication
  • Dynamic HMI generation for in-vehicle infotainment

Research assistants in this area:

A major focus of our research activities is in the area of mobile multimedia communication. Wireless digital multimedia communication faces additional challenges when compared to wired multimedia communication. In particular, dynamic resource allocation in multi-user wireless systems is currently on our research agenda. This is challenging due to the time-varying characteristics of wireless links and the limited adapatability of multimedia applications. Cross-layer design and optimization is a promising direction to address these challenges. Recent topics are:

  • Joint source channel coding
  • Cross-layer design and optimization
  • Utility-based resource allocation in multi-user wireless networks
  • Source and loss distortion modeling of audio-visual applications
  • Error-resilient video encoding and transmission
  • Application-driven scheduling and resource allocation in wireless systems 

Research assistants in this area:

A major focus of our research activities is in the area of three-dimensional image analysis and synthesis. The former is traditionally studied in the computer vision community. The latter has been studied mainly by the computer graphics community. Computer vision, computer graphics and image/video commmunication are no longer separate fields. Today, major innovations happen at the intersection of these traditionally independent areas. Recent projects/activities in this area address for example:

  • Spatial and temporal synchronization of video sequences
  • Feature detection, matching, and tracking
  • Visual localization, odometry and mapping
  • Mobile Augmented Reality
  • Acquisition of joint geometry and image based 3D worlds
  • Sensor calibration and sensor data fusion
  • Surprise detection using image based scene representations