Advanced Seminar / Masterseminar SS15
The seminar will have only few fixed class meetings. Besides these meetings, additional appointments shall be arranged individually and on demand. The schedule of this seminar can be found in the online booking system (OBS).
Each group’s term paper has to be prepared using Springer’s LNCS template with a length of 10 – 12 pages without references and clear marked appendices. The LNCS templates for Word and LaTeX can be found here: http://www.springer.com/computer/lncs?SGWID=0-164-6-793341-0. The final presentation will be 40 minutes per group + 20 minutes discussion of the results. A grade will be given based on the term paper and final presentation as defined in the module description.
According to OBS, the course capacity is 12 seats. Eventually, additional seats can be offered. If you don’t get a seat in the allocation period, please join the kickoff and ask for an additional seat. If you got allocated a seat but can’t join the kickoff, please get in touch with me by email before the kickoff. Otherwise, your seat will be revoked.
Student groups (1-2 persons) may select their own topic or choose a topic from the following list:
- Survey on Ageing Effects on Biometric Recognition Systems:
Based on a thorough literature study on biometric ageing with respect to fingerprints, iris and vein biometrics students should compose a comprehensive survey summarizing, comparing and discussing results of existing works in the field (2010-2015). For this topic the term paper should be at least 16 pages.
- SIFT-based Correlation Attacks on Cancelable Biometric Systems:
Cancelable biometrics should provide higher security with respect to the storage of biometric data. However, SIFT-based correlation attacks may be feasible. For this topic students should consider cancelable biometric systems in the image domain. Students will implement a simple block-based permutation scheme which will then be attacked employing an open-source implementation of the SIFT operator.
- Extension of Open-Source Iris Recognition Software:
USIT is an open-source iris recognition toolkit which should be extended implementing a new feature extraction module. The feature extraction should either be a re-implementation of „Ordinal Measures“ or „2D-Gabor Filters“. In case of successful implementation and testing the module will be integrated to the software toolkit.
- Extracting robust i-vector Features in Speaker Recognition:
Noise and duration variations are known to cause biometric performance breakdowns in speaker recognition. For the purpose of extracting more robust intermediate-sized vector (i-vector) features, Vector Tailor Series (VTS) and Deep Learning based i-vector estimations promise to yield sufficient gains. Students will implement both approaches, and discuss pros/cons towards a State-of-the-Art baseline system.
- Effects of varying Audio Compression Formats to Voice Biometrics:
Lossy audio compression reveals lesser sample quality for extracting proper voice patterns. The term paper should cover a comparison of the most common audio compression formats for speech and video files, and examine their effects on a given speaker recognition system.
- Survey on Voice Activity Detection Algorithms:
Sample segmentation is an important signal processing step for biometric comparisons. Based on a literature study on voice activity detection (VAD) algorithms in speech recognition, a term paper should be derived comparing and discussing VAD approaches. The term paper should cover an experimental evaluation of the most promising VAD techniques and their impacts on a speaker recognition system as well. For this topic the term paper should be at least 16 pages.