The main area of research and teaching activities carried out by the employees of the Department are issues correlated with biometrics, signal processing and analysis, in particular sound and image, using traditional methods as well as machine learning and artificial intelligence tools.
Another important area is the work related to digital technology and embedded systems, which are key in the Internet of Things industry. The whole is completed by the issues of human-computer communication, modern NUI interfaces (Natural User Interface) and Augmented Reality. The above-mentioned issues are very important elements of the Industry of the Future, also known as Industry 4.0.
Research topics conducted at the Department of Digital Media and Computer Graphics focuses on several main trends:
- Biometrics; Recognition of people based on the methods of analyzing biometric data, body parts and signals collected by various devices;
- Analysis of medical images (including the detection of lesions at various stages). Artificial intelligence and machine learning techniques in image classification;
- Digital signal processing and analysis; audio signal processing techniques; object detection and classification in image processing;
- Speech processing, with particular emphasis on the processing and coding of the speech signal; Signal decomposition techniques and block transforms;
- Use of machine learning tools in the processing of audio and image signals.
The topics and scope of research of individual employees include:
- Expert systems. Knowledge representation; Processing uncertain knowledge. Inference methods under uncertainty;
- Embedded systems on programmable circuits, finite state machine synthesis, design of control systems, high-level design of digital systems;
- Robotics; Identification of objects and surroundings; Topological navigation algorithms based on an integrated data collection system;
- Applications of image processing and recognition techniques, incl. automatic recognition of biomarker concentration (SPRi), recognition of dangerous objects in X-ray scanner images, recognition of objects on the road of the car;
- The issues relate, inter alia, to classification of objects using texture features, R-CNN network, key point detection and segmentation algorithms.
- Signal processing with the use of quaternions algebra and deep learning with particular emphasis on the series of time sensors of motion;
- Application of machine learning algorithms, including deep learning, to automatically characterize spatial audio recordings;
- Speech signal analysis using distorted transforms and signal subspace methods; Single-channel speech signal conditioning systems; Estimating noise in acoustic signals; Use of psychoacoustic phenomena in the processing of the speech signal;
- Use of fuzzy random forests based on cluster models – fuzzy decision trees; Gesture recognition and issues related to drones and image analysis;
- Designing control systems, in particular sequence systems based on programmable CPLD / FPGA systems;
- Methods of logical synthesis, optimization of the footprint, speed of operation and consumed power;
- Machine learning tools, cloud technologies and Cognitive Services in the process of codifying organizational knowledge and knowledge management;
- Identification of dynamic systems: continuous, discrete, linear and nonlinear. Selection of optimal and friendly control signals in the task of identifying dynamic systems;
- Technical information technology and telecommunications, with particular emphasis on the design of embedded systems and the design of digital circuits based on programmable CPLD / FPGA circuits;
- “Information technology and telecommunications”, with particular emphasis on behavioral biometrics and the use of IMU (Inertial Measurement Unit) measuring sensors.