The Multimedia Pattern Recognition research group works in one of the six major research areas of Fraunhofer IAIS. We develop advanced algorithms for the segmentation, recognition, and indexing of speech, audio, image, and video data. We also analyze social media and investigate how people interact in a networked world.

Please visit our scientific blog for more information.

Latest Blog News

We've now got an open source implementation of the Region Covariance descriptor, accompanied by some example code, all based on OpenCV.

Our paper on network growth and the spectral evolution model has been accepted as a full paper at the ACM International Conference on Knowledge and Information Management.

Our paper on simplex volume maximization for descriptive, web-scale matrix factorization has been accepted for publication at the ACM International Conference on Information and Knowledge Management.

Pattern Recognition Letters have dedicated a special issue on "Pattern Recognition of Non-Speech Audio". In that issue, we present an interdisciplinary collaboration on "Detecting bird sounds in a complex acoustic environment and application to bioacoustic monitoring", together with researchers from the University of Bonn, Fraunhofer FKIE, and the Animal Sound Archive at the Humboldt University, Berlin.

Our paper on contextual verification was accepted for Interspeech 2010! The submission was prepared in collaboration with my colleagues Timo Mertens from NTNU Trondheim and Martha Larson from TU Delft. See you in Japan!