||Information Extraction from Handwritten Notes
- Overview and Future Perspective of Handwriting Technologies Marcus Liwicki (DFKI)
Are handwritten fated to disappear due to the huge popularity of keyboards and their very high efficiency? Even the newest market shares of pen-enabled devices are still quite low when compared to devices with on-screen keyboards. However, this development will change drastically with the new evolution of applications which clearly show the strength of handwritten input. The most typical example of using handwriting is note-taking. In this presentation I will present several technologies which make it able to finally understand the handwritten input and extend the digitized knowledge on the user's PC. These technologies include handwriting recognition and information extraction, which will be reviewed and the state-of-the-art will be presented. Next, an application for question-answering will be described and presented. Finally, the future perspectives of handwriting and handwritten input technologies will be presented. My hypothesis is that the question above can be answered with a clear NO. In contrast, due to the increasing need for intuitive input devices, the keyboard and mouse will more and more disappear. Novel devices with high-end applications demonstrate the usefulness of the pen and the handwriting recognition technology is ready to be integrated as well.
Embodied Interaction - Projects and Perspectives Markus Weber (DFKI)
Since devices for body-motion tracking are now available on the consumer market, Embodied Interaction techniques become more and more important. The Virtual Physiotherapist system of the PAMAP (Physical Activity Monitoring for Aging People) project offers monitored strength exercises using personalized models. Using a sensor network with a configurable inertial measurement unit (IMU) setup captures the human motion. The system counts the repetitions and sets and evaluates exercise load and technique taking speed, amplitude, smoothness, and posture into consideration. Based on this, the monitored person gets immediate feedback (e.g., by raised alarms or messages) as well as graphical help to perform the correct motion. The captured movements are visualized by a virtual character. Temporal Pattern Recognition techniques are utilized to cope with issues such unsupervised motion model learning and online real-time motion segmentation.
Another interesting domain for the usage of Embodied Interaction is in factory training. The VISTRA (VIrtual Simulation and TRAining) project aims at the development of a comprehensive platform for simulation and training of manual assembly processes based on advanced ICT-technologies and concepts, such as auto-generation and re-use of data, realistic physical behavior, game-based learning, advanced user-interaction and cross-disciplinary information sharing. Having factory workers using embodied interaction during the learning process seems to be a promising approach to improve the training process.