A list of available thesis topics can be found on the Slovenian version of this page. Foreign students are invited to the lab for a discussion on possible topics. In general the members of the lab cover different topics related to artificial intelligence, pattern recognition, machine learning, biometrics, speech technologies, computer vision, deep learning, and other related areas.
Below are some topic ideas.
- Face recognition (contact: Vitomir Štruc)
- Developing a face recognition engine
- Developing a 3D face recognition engine
- Comparative assessment of photometric normalization techniques
- Comparative assessment of face detection and/or landmark localization techniques
- Local descriptors for face recognition
- Developing a demo face morphing app
- Iris recognition (contact: Vitomir Štruc)
- Developing a iris recognition engine
- Building a iris segmentation (ROI extraction) module for iris recognition
- Comparative assessment of preprocessing techniques on iris recognition
- Palm, finger and hand-geometry recognition (contact: Vitomir Štruc)
- Developing the recognition engine
- ROI detection and extraction
- Contactless recognition based on palms, fingers or hand geometry
- Using 3D information in palm recognition systems
- Vein-based personal recognition (contact: Vitomir Štruc)
- Developing the recognition engine
- Various comparative assessments
- Speech recognition (contact: Simon Dobrišek)
- Developing a Sphinx module for robust speech detection
- Spatial-localization of speakers based on accustic signals
- Adaptation of grapheme-to-phoneme conversion for slovenian speech
- Language modeling of out-of-dictionary words for the slovenian speech recognition engine
- Extraction and analysis of a word-dictionary from a larger corpus
- Speaker recognition (contact: Simon Dobrišek)
- Collection and analysis of a speech corpus for a speaker recognition engine capable of operating in ambient intelligence environments
- Developing a robust speaker recognition engine
- Spatial speaker localization based on speech
- Recognition of suspicious sounds for surveillance applications (contact: Simon Dobrišek)
- Collection and analysis of a corpus with suspicious sounds
- Accustics-based spatial localization of suspicious sounds
- Developing a recognition engine for shouts and cries for help
- Developing a recognition engine for suspicious sounds
- Controling PTZ cameras and directing them towards the sources of suspicious sounds
- Speech synthesis (contact: Simon Dobrišek)
- Automatic alignment of speech transcriptions on slovenian speech corpora
- Automatic transcription-error detection based on slovenian speech corpora
- Assessment of parameter impact in HMM-based slovenian speech synthesizers