Image and Video Analysis

Image and Video Analysis

The Image and Video Analysis (IVA) team is a research team within IVML Laboratory at National Technical University of Athens. It focuses on basic research and applications of computer vision, machine learning and information retrieval to analyze, interpret, and search static and moving images (video).

The team studies and develops methods for local feature and salient region detection, spatial and spatiotemporal visual representation and matching, object recognition and tracking, action recognition in video, scene classification, image/video indexing, retrieval and summarization. The analysis process often includes information from other modalities like text and numeric data.

The latest achievements of the team include highly repeatable feature detection, extremely fast spatial matching, geometry indexing, multi-view and single-view feature selection, approximate nearest neighbor search for large scale clustering and vocabulary construction, and mining 3d scenes from millions of images. Most implementations are based on our generic C++ library ivl.

Applications include web image/video search, organization of personal and community photo collections, video exploration, visual surveillance, industrial visual inspection, and indexing of image/video archives. To get an idea, see our unique application VIRaL.

News
26.03.2015 Approximate nearest neighbor search: binary codes and vector quantization
Invited talk at University of Athens [Slides]
02.10.2014 Image retrieval, vector quantization and nearest neighbor search
Invited talk at INRIA Rennes [Slides]
27.04.2014 Paper at CVPR 2014: Locally Optimized Product Quantization for Approximate Nearest Neighbor Search
04.01.2014 Paper in CVIU: large-scale geometry indexing
04.01.2014 Aggregated Selective Match Kernel (ASMK): code available
22.12.2013 Dimensionality-recursive vector quantization (DRVQ): C++ source code available
17.12.2013 Advanced projects for Image and Video Analysis course: list of topics, bibliography
02.10.2013 Papers at ICCV 2013: to aggregate or not to aggregate and quantize and conquer
and IJCV: Hough pyramid matching
26.06.2012 Papers at ECCV 2012: WαSH detector and approximate Gaussian mixtures
and ACM Multimedia 2012: SymCity
22.06.2012 ivl 0.9a is out: a template C++ library extending syntax towards mathematical notation
Visit ivl directly at http://image.ntua.gr/ivl/
14.06.2012 VIRaL goes mobile: Android app FindmyPhoto is now available at Google play
11.12.2011 VIRaL 2.0 is now available, featuring Explore and Routes
06.11.2011 Papers at ICCV 2011: Medial Features and Relaxed Spatial Matching
31.03.2011 Image matching and visual search: local features and geometry
Invited talk at University of Bordeaux [Slides - 92MB]
20.10.2010 Presentation of Diploma Theses for 2010-2011 [Handout] [Slides]

Contact
Email: iavr@image.ntua.gr
Phone: +30 210 772 3038
Fax: +30 210 772 2492
Address: Iroon Polytexneiou 9, 15780 Zografou, Greece
Electrical Engineering Building, 1st Floor - Room 1.1.23