Kolloquium SoSe 2017

Speaker

Prof. Dr.-Ing. Kai Uwe Barthel

Institution: Hochschule für Technik und Wirtschaft, HTW Berlin

Internationaler Studiengang Medieninformatik

Webseite: www.visual-computing.com

When: 29.05.2017

Where: Room B-201, Informatikum

Topic

Visually Browsing Millions of Images using Image Graphs

Abstract

In the past an efficient and satisfactory image search was only possible by using a combation of keywords and low-level visual image features. Recently Convolutional Neural Networks (CNNs) have enabled automatic understanding of images. This results in a multitude of new applications and improved visual image search systems. This talk provides an overview of the different methods for image search, gives an overview of the principle of CNNs and shows how future image search systems could look like. We present a new approach to visually explore very large sets of untagged images. High quality image descriptors are generated using transformed activations of a convolutional neural network. These features are used to model image similarities, from which a hierarchical image graph is build. We show how such a graph can be constructed efficiently. Best user experience for navigating this graph is achieved by projecting sub-graphs onto a regular 2D-image map. This allows users to explore the image graph similar to navigation services.

Vita

Prof. Dr. Kai Uwe Barthel studied Electrical Engineering at the Technical University of Berlin. At the university’s department for image and signal processing he finished in 1996 his PhD thesis about fractal image compression. In 1998 he became head of R&D with LuraTech Inc. where hard- and software for image compression were developed. He also participated in the JPEG2000 standardization committee. His research topic of that time was the segmentation of documents for Mixed-Raster-Content compression. In 2001 he became a professor for visual computing at HTW Berlin, University of Applied Sciences. At HTW Berlin he is teaching courses such as image processing, computer vision, visual information retrieval, and machine learning. Main research topics are visual image search and automatic image understanding. In 2009 he founded pixolution, a company for visual image search. His latest research includes automatic keywording of images, image clustering and the development of visual image navigation systems such as www.picsbuffet.com

Host

Prof. Dr.-Ing. H.-Siegfried Stiehl, BV/HCC