prof. RNDr. Tomáš Skopal, Ph.D.

The complexity of next-generation retrieval systems originates from the requirement to organize massive and ever growing volumes of heterogeneous data and meta-data, together with the need to provide distributed management prevalently based on similarity matching. The problem starts with data acquisition of weakly structured or completely unstructured data, such as images and video, which necessarily need innovative techniques for information extraction and classification to increase their findability. In principle, we consider the object findability and the actual search process as two fundamental and synergic aspects of the retrieval. Both of them pose effectiveness and efficiency challenges which need innovative theories and technologies, and must be studied together to converge to qualitatively new retrieval tools of the future. The dissertation topic is foundational in nature as it addresses the theoretical limits of similarity retrieval in context of the Big Data problem. Fundamental to the thesis is the development of scalable solutions.

prof. RNDr. Tomáš Skopal, Ph.D.

RNDr. David Hoksza, Ph.D.

Department of Software Engineering

prof. RNDr. Tomáš Skopal, Ph.D.

Ing. Jan Baier

Department of Software Engineering

The aim of this work is to design a method of efficient HRTF acquisition from the profiles already existed. The thesis analyzes the possibilities of using the recommender systems to find a suitable HRTF that could replace the most commonly used approaches. Using web applications - implemented by technologies WebGL and WebAudio we have verified the effectiveness of collaborative filtering. User testing demonstrated improvement of accuracy against an unpersonalized HRTF profile of an averagely 10 degrees. The conclusion of this work is to find a fast method for relatively accurate audio virtual reality simulation.

Thesis on DSpace

prof. RNDr. Tomáš Skopal, Ph.D.

Department of Software Engineering

This master's thesis is dedicated to the design and implementation of web interface for real-time video analytics system named Videolytics. Main target of this system is extraction, storage and processing of high-level features from video surveillance data. The resulting application is able to combine various data sources and modules: video stream, object detections and trajectories from database, management of processes for data generation. Developed application touches on various different aspects: video stream management, effective database querying, precise (milliseconds range) synchronization of data sources, server process management. Maintaining large number of different aspects inside one application is highly complex. This is why Videolytics web portal is based on microservices architecture. Web client communicates with services using HTTP protocol, while each service accomplishes it's certain distinct goal. Web portal collaborates with other modules implemented in different programming languages. All of them are functioning on single Unix server to avoid redundant network load. Resource sharing and dependencies are resolved using docker virtualization technology. Web portal is implemented using Apache web server in combination with PHP language.

Thesis on DSpace