J. Gelšvartas “Multifunctional user interface for quadriplegics: development and investigation” doctoral dissertation defence

Author, Institution: Julius Gelšvartas, Kaunas University of Technology

Science area, field of science: Technological Sciences, Informatics Engineering, T007

Scientific Supervisor: Prof. Dr. Habil. Rimvydas Simutis (Kaunas University of Technology, Technological sciences, Informatics engineering, T 007).
Scientific Advisor: Dr. Rytis Maskeliūnas (Kaunas University of Technology, Technological
sciences, Informatics engineering, T 007).

Dissertation Defence Board of Informatics Engineering Science Field:
Prof. Dr. Habil. Antanas Verikas (Kaunas University of Technology, Technological sciences, Informatics engineering, T 007) chairman,
Dr. Lalit Garg (Malta University, Technological sciences, Informatics engineering, T 007),
Prof. Dr. Arnas Kačeniauskas (Vilnius Gediminas Technical University, Technological sciences, Informatics engineering, T 007)
Prof. Dr. Olga Kurasova (Vilnius University, Technological sciences, Informatics engineering, T 007)
Assoc. Prof. Dr. Evaldas Vaičiukynas (Kaunas University of Technology, Technological sciences, Informatics engineering, T 007).

The doctoral dissertation is available at libraries of Kaunas University of Technology (Donelaičio 20, Kaunas) and Vilnius Gediminas Technical University (Saulėtekio al. 14, Vilnius).

Annotation:

This dissertation presents a multifunctional user interface developed for people suffering from severe motor disabilities. The presented user interface can be used to input text, select real world objects and answer questions. Dissertation presents a detailed architecture of the developed system. Dissertation also presents novel algorithms that help to ensure system efficiency. Projection mapping user interface automatically detects tabletop objects and highlights them using projector. Efficient camera control algorithm helps to ensure the stability of computer vision algorithms. We present a cascade-based camera parameter control algorithm with automatic gain scheduling. Dissertation also contains a description of region based stereo matching algorithm. The algorithm is capable of processing images in real time on an embedded computer. Text input user interface uses a specialized text prediction algorithm that decreases text input time. Experiments were performed to demonstrate the benefits of the proposed algorithms. Developed user interface was evaluated using participant questionnaire and measuring task execution times.