This is "The Multimedia Signal Processing and Security Lab", short WaveLab, website. We are a research group at the Artificial Intelligence and Human Interfaces (AIHI) Department of the University of Salzburg led by Andreas Uhl. Our research is focused on Visual Data Processing and associated security questions. Most of our work is currently concentrated on Biometrics, Media Forensics and Media Security, Medical Image and Video Analysis, and application oriented fundamental research in digital humanities, individualised aquaculture and sustainable wood industry.
Media Security, Media Forensics & Watermarking
In the area of media security we are specializing in lightweight and partial encryption schemes for image and video data
with special emphasis on scalable media like JPEG2000 or H.264 SVC, and the assessment of the visual security such schemes do offer.
In media forensics, we have been working on the usage of PRNU-based sensor authentication techniques in a biometric centext, on the detection
of morphed portraits and deep fakes, and on robustness of copy-move forgery detectors.
In the watermarking area our focus is on
developing bitstream-based embedding techniques and key-dependency schemes for robust embedding techniques. Scalable watermarking and multiple watermarking are recent topics
of reasearch. In robust hashing, key-dependent wavelet transforms are also investigated as a means to provide security for such
We focus on biometric modalities where image processing is conducted during feature extraction and
template generation (i.e. vascular biometrics like finger veins or hand veins, fingerprints, iris, face, retina, hand and foot geometry).
Current emphasis of our work
is on NIR sensor construction, sample data compression and encryption, as well as on privacy (like cancelable biometrics) and bias in biometric systems.
Questions of recognition robustness play an important in our research, and we also consider biometrics as applied e.g. in tree tracing and fish identification.
Medical Image Analysis and Classification
Together with partners from different medical departments, we apply visual analysis and classification techniques to medical image data
targeted towards the development of computer-based decision support systems. Target imaging modalities so far have been gestro-intestinal endoscopic imagery
and (brain) MRI. On the one hand,
classical model-based texture classification features are developed and used, on the other hand data-driven learning based schemes are employed and customized to the
target imagery. We have been working on a staging technique for colon cancer, diagnosis of celiac disease, and discrimination of mild cognitive impairment from
epilepsy based on MRI.
Image- and Videocoding
We have been specializing in image and video compression techniques for over 25 years now.
In the development of techniques, emphasis has been given to adaptive techniques like wavelet packets or object based coding inspired by MPEG-4.
Also, cache efficient and parallel algorithms have been developed for a number of corresponding algorithms.
Often, we concentrate on scalable coding schemes like JPEG2000, MCTF scalable wavelet codecs, and H.264 SVC. More recently, we have been assessing the
employment of various lossy compression techniques in specific application contexts, like in biometric recognition, in media forensics and security, and in microscopy imaging.
General Computer Vision and Machine Learning Topics
Of course, we have also been involved in more general Computer Vision and Machine Learning topics, mostly collaborating either with companies, or supporting other research groups
not involved in visual data processing.