November 2009 - today
Image Retrieval using Scene Completion
The initial query image is of paramount importance when searching for images of interest, because it has a direct effect on the retrieval results. Even though the image that is chosen as the query naturally is similar to what the user is looking for, it nonetheless often contains elements that are irrelevant to what the user has in mind. In this project we develop a novel image retrieval system where the user can remove such unwanted image elements and seamlessly fill in the resulting holes with semantically meaningful and desired content. This composite image is then used as the actual query image.

July 2009 - June 2010
TOP-SURF: a visual words toolkit
TOP-SURF is an image descriptor that combines interest points with visual words, resulting in a high performance yet compact descriptor that is designed with a wide range of content-based image retrieval applications in mind. In addition to the source code for the visual word extraction and matching, we also provide a high level API and very large pre-computed codebooks targeting web image content for both research and teaching purposes.

January 2009 - October 2009
Interactive Visual Exploration of Image Collections
Experiential image retrieval systems aim to provide the user with a natural and intuitive search experience. The goal is to empower the user to navigate large collections based on his own needs and preferences, while simultaneously providing her with an accurate sense of what the database has to offer. In this project, a novel interface gives the user the opportunity to visually and interactively explore the feature space around relevant images and to focus the search on only those regions in feature space that are relevant.

January 2008 - February 2010
Image Copy Detection
This project provides a comparative study of content-based copy detection methods, which include research literature methods based on salient point matching, discrete cosine and wavelet transforms, color histograms and a biologically motivated visual matching method. The evaluation focuses on large-scale applications, especially on performance in the context of search engines for web images. For our experiments, original images have been altered by a diverse set of realistic transformations and have been embedded in a collection of one million web images and one million Flickr images.

August 2007 - present
Noteworthy Image Search
This text-based internet search engine finds the newest images for any particular query, especially for noteworthy imagery, i.e. "what are the newest images for 'Star Wars' since my last visit?". To do this, we keep track of images as they are downloaded, cluster them by similarity and determine by date which ones are new. The similarity clustering is done by using descriptive keywords extracted from the page the image was found on and by using content features extracted from the image.

September 2006 - December 2008
Visual Information Retrieval Using Synthetic Imagery
Funded by NWO and BRICKS
Our own visual imagination allows us to form mental images based on our memories and experiences. When we are learning new visual concepts, we often construct such images based on real objects or scenes to help understand or clarify the primary features which are associated with the concept. A new paradigm we call artificial imagination is the digital analogy of our own visual imagination. By endowing the retrieval system with this capability the search results are likely to improve.

January 2006 - April 2007
Immersive and Spatial Voice Audio in Networked Virtual Environments, a.k.a. "The Matrix"
In the current work on multiplayer virtual reality, the research has focused largely on the visual aspect combined with text. Audio is often neglected and when present typically ignores the structure of the virtual world. In this project we have created a new system which integrates the audio, visual and 3D structure of the virtual world. Specifically, our novel contribution is the creation of a system which models the effect of the 3D world structure upon the audio and visual aspects in a natural and intuitive manner: players in the massive multiplayer world can now talk with each other as in real life.

July 2004 - August 2005
Video Analysis and Classification using MagicVideoBrowser
Built specifically for movie enthusiasts, MagicVideoBrowser offers a highly intuitive interface that allows its users an easy and effective way of navigating one's computer to browse or search for video files. Videos are represented in a truly visual way as opposed to text: the visual elements are carefully selected by plugins that analyze the video content and use classification methods to select the - in their view - most important scenes. These scenes are then used as a representative view of the video.

February 2005 - July 2005
Pandora's Digital Box
Intrigued by the ongoing debate about privacy, the constant confrontation with viruses and worms and the abundance of advertisements, this project was created to bring the false sense of safety that surrounds mobile equipment under attention. A device, equipped with GSM, Bluetooth and internet hotspot capabilities, showed that it's not the technological barrier that stops people with evil intentions from intruding on us, but that currently the law is our only protection.

January 2004 - June 2004
Dance Fantasia
This project aims to create a dancing environment, where the visual, sound and motion input and output synchronize as one. The dancing atmosphere is represented in terms of natural elements: water, body and fire. Real-time video capture of dancing people combined with the beat of the music control the mixed-down video output: by dancing, the people determine which video footage is played, and their movements to the music lead to the drawing of elements on the screen. The elements change position to follow the user and change size according to the speed of the user's movements.