The research in the Signal Processing and Recognition Group concerns theory, algorithms and systems for analyzing, understanding, classifying and predicting physical processes based on data/signals in the following seven overlapping and interacting theoretical and applied themes:
- Audio, image and video processing, encompassing compression, restoration, enhancement, segmentation and identification tasks;
- Pattern recognition, including representation, dimensionality reduction and generalization issues;
- Bioengineering, cooperating with medical doctors mainly in language therapy, electrocardiography, electroencephalography and medical images.
- Agro-industrial applications, cooperating with experts in fields such as plant pathology, flavor analysis and plant bioinformatics;
- Volcanic seismology, cooperating with seismologists for analyzing seismic signals and automate their recognition;
- Industrial inspection, including damage prognosis and health monitoring of civil structures and mechanical parts;
- Motion analysis by applying manifold learning techniques;
- Other applications such as meteorology and soil dynamics.
The group collaborates with a number of academic and industrial partners, such as:
As a result of the research projects, the Group has built and registered the following data sets (available upon request):