A Recurrent Neural Network for Hand Gesture Recognition based on Accelerometer Data.

For many applications, hand gesture recognition systems that rely on biosignal data exclusively are mandatory. Usually, theses systems have to be affordable, reliable as well as mobile. The hand is moved due to muscle contractions that cause motions of the forearm skin. Theses motions can be captured with cheap and reliable accelerometers placed around the forearm. Since accelerometers can also be integrated into mobile systems easily, the possibility of a robust hand gesture recognition based on accelerometer signals is evaluated in this work. For this, a neural network architecture consisting of two different kinds of recurrent neural network (RNN) cells is proposed. Experiments on three databases reveal that this relatively small network outperforms by far state-of-the-art hand gesture recognition approaches that rely on multi-modal data. The combination of accelerometer data and an RNN forms a robust hand gesture classification system, i.e., the performance of the network does not vary a lot between subjects and it is outstanding for amputees. Furthermore, the proposed network uses only 5 ms short windows to classify the hand gestures. Consequently, this approach allows for a quick, and potentially delay-free hand gesture detection.

Assembling proteomics data as a prerequisite for the analysis of large scale experiments.

BACKGROUND: Despite the complete determination of the genome sequence of a huge number of bacteria, their proteomes remain relatively poorly defined. Beside new methods to increase the number of identified proteins new database applications are necessary to store and present results of large- scale proteomics experiments. RESULTS: In the present study, a database concept has been developed to address these issues and to offer complete information via a web interface. In our concept, the Oracle based data repository system SQL-LIMS plays the central role in the proteomics workflow and was applied to the proteomes of Mycobacterium tuberculosis, Helicobacter pylori, Salmonella typhimurium and protein complexes such as 20S proteasome. Technical operations of our proteomics labs were used as the standard for SQL-LIMS template creation. By means of a Java based data parser, post-processed data of different approaches, such as LC/ESI-MS, MALDI-MS and 2-D gel electrophoresis (2-DE), were stored in SQL-LIMS. A minimum set of the proteomics data were transferred in our public 2D-PAGE database using a Java based interface (Data Transfer Tool) with the requirements of the PEDRo standardization. Furthermore, the stored proteomics data were extractable out of SQL-LIMS via XML. CONCLUSION: The Oracle based data repository system SQL-LIMS played the central role in the proteomics workflow concept. Technical operations of our proteomics labs were used as standards for SQL-LIMS templates. Using a Java based parser, post-processed data of different approaches such as LC/ESI-MS, MALDI-MS and 1-DE and 2-DE were stored in SQL-LIMS. Thus, unique data formats of different instruments were unified and stored in SQL-LIMS tables. Moreover, a unique submission identifier allowed fast access to all experimental data. This was the main advantage compared to multi software solutions, especially if personnel fluctuations are high. Moreover, large scale and high-throughput experiments must be managed in a comprehensive repository system such as SQL-LIMS, to query results in a systematic manner. On the other hand, these database systems are expensive and require at least one full time administrator and specialized lab manager. Moreover, the high technical dynamics in proteomics may cause problems to adjust new data formats. To summarize, SQL-LIMS met the requirements of proteomics data handling especially in skilled processes such as gel-electrophoresis or mass spectrometry and fulfilled the PSI standardization criteria. The data transfer into a public domain via DTT facilitated validation of proteomics data. Additionally, evaluation of mass spectra by post-processing using MS-Screener improved the reliability of mass analysis and prevented storage of data junk.