gms | German Medical Science

Introduction: Biomedical signals and images are observations of physiological processes and their impact for healthcare applications is still rising. Examples are time series like ECG, EEG or intracortical multi-electrode recordings. At the other end of the spectrum we have image data recorded by microscopy, magnetic resonance imaging (MRI), computed tomography (CT) or ultrasound systems. Of particular importance are combinations of various measurement methods and the subsequent processing of the high dimensional data sets. Processing of signals as well of images is divided into several methodological parts like pre-processing, noise reduction, fusion, analysis and pattern recognition. Although the requirements for biomedical image and signal processing are very similar, both have developed to different and autonomous disciplines of research and development. The main goal of the newly established workshop Biomedical Image and Signal Processing (BISC) is to bridge the gap between image and signal processing. We aim to enhance the methodological exchange, identify areas of overlap, find new trends especially at the edge between both areas and foster personal communication across disciplines and special interest groups. We thank especially the following associations for their support:

• Deutsche Gesellschaft für Med- Informatik, Biometrie und Epidemiologie (GMDS)
• Gesellschaft für Informatik (GI)
• Deutsche Gesellschaft für Biomedizinische Technik im VDE (DGMT).

Results: The content range of accepted abstracts is wide in terms of information sources, focus of application as well as methodology. Images and signals result from MRI, functional MRI, x-ray, microscopy, ultrasound, CT and ECG. Methods to deal with these information sources start with pre-processing, semi-classical signal analysis and signal analysis in the time-frequency domain. They are carried forward with texture analysis, reconstruction and filtering, visualization and parallelization to end up with segmentation, classification and registration methods. Some approaches are formulated for general purpose applications in the biomedical domain, some others are developed specifically for diseases like brain tumors, breast cancer, major depressive disorder and gastro-esophageal reflux disease or for body parts like lungs or ventricles. Interestingly, the majority of abstracts focus on combinations of different disciplines, methods or imaging modalities. The processing of ultrasound images interpreting the time course as signal results in a spatio-temporal analysis. The application of image processing filters to the visualization of ECG signals combines image and signal processing methods for improved analysis. The combination of registration and segmentation methods aims to benefit from both methods. In addition, data recorded at different points in time allow follow-up diagnosis and novel time-course analyses. The fusion of different imaging modalities provides more insight into the relationship of structures and, thus, improves diagnosis. Hence, the BISC workshop - combining the disciplines biomedical image and signal processing - points in the right direction where innovations can be expected.