- Video Analysis and Summarization
- Advanced Image and Video Retrieval Techniques
- Image Retrieval and Classification Techniques
- Multimodal Machine Learning Applications
- Music and Audio Processing
- Advanced Vision and Imaging
- Speech and dialogue systems
- Human Pose and Action Recognition
- Data Stream Mining Techniques
- Peer-to-Peer Network Technologies
- Advanced Database Systems and Queries
- Data Management and Algorithms
- Anomaly Detection Techniques and Applications
- Video Coding and Compression Technologies
- Natural Language Processing Techniques
University of Basel
2012-2019
vitrivr is an open source full-stack content-based multimedia retrieval system with focus on video. Unlike the majority of existing search solutions, not limited to searching in metadata, but also provides and thus offers a large variety different query modes which can be seamlessly combined: Query by sketch, allows user draw sketch image and/or motion paths, example, keyword search, relevance feedback. The architecture self-contained addresses all aspects from offline feature extraction,...
Despite the tremendous importance and availability of large video collections, support for retrieval is still rather limited mostly tailored to very concrete use cases collections. In image retrieval, instance, standard keyword search on basis manual annotations content-based based similarity query (s), are well established paradigms, both in academic prototypes commercial engines. Recently, with proliferation sketch-enabled devices, also sketch-based has received considerable attention. The...
Recent developments in sport analytics have heightened the interest collecting data on behavior of individuals and entire team sports events. Rather than using dedicated sensors for recording data, detection semantic events reflecting a team's subsequent annotation video is nowadays mostly performed by paid experts. In this paper, we present an approach to generating such annotations leveraging wisdom crowd. We CrowdSport application that allows collect soccer games. It presents crowd...
The digitization of museum exhibits has raised the question how to make these data accessible, particularly in light ever growing collections being available. In this demo, we present VIRTUE system which allows curators easily set up virtual exhibitions static and dynamic 2D (paintings, photographs, videos, etc.) 3D artifacts. Visitors may navigate through rooms, inspect artifacts interact with them novel ways. Participants will be able use by creating their own exhibitions, they tour as a visitor.
The past decade has seen the rapid proliferation of low-priced devices for recording image, audio and video data in nearly unlimited quantity. Multimedia is Big Data, not only terms their volume, but also with respect to heterogeneous nature. This includes variety queries be executed. Current approaches searching big multimedia collections mainly rely on keywords. However, manually annotating every single object a large collection feasible. Therefore, content-based retrieval -using sample...
The IMOTION system is a multimodal content-based video search and browsing application offering rich set of query modes on the basis broad range different features. It able to scale with size collection due its underlying flexible polystore called ADAMpro very effective retrieval engine Cineast, optimized for multi-feature fusion. simultaneously geared towards precision-focused searches, i.e., known-item image or text queries, recall-focused, exploratory searches. In this demo, we will...
The IMOTION system is a content-based video search engine that provides fast and intuitive known item in large collections. User interaction consists mainly of sketching, which the recognizes real-time makes suggestions based on both visual appearance sketch (what does look like terms colors, edge distribution, etc.) semantic content object user sketching). latter enabled by predictive sketch-based UI identifies likely candidates for sketched via state-of-the-art recognition techniques...
The enormous increase of digital image collections urgently necessitates effective, efficient, and in particular highly flexible approaches to retrieval. Different search paradigms such as text search, query-by-example, or query-by-sketch need be seamlessly combined integrated support different information needs allow users start (and subsequently refine) queries with any type object. In this paper, we present QUEST (Query by Example, Sketch Text), a novel multi-modal content-based retrieval...
This paper presents an after-the-fact summary of the participation vitrivr system to 2019 Video Browser Showdown. Analogously last year's report, focus this lies on additions made since original publication and system's performance during competition.
URL: https://vitrivr.org/ vitrivr is an open source retrieval system capable of processing multimedia documents such as images, videos, music, and 3D-models. It supports a wealth content based features for multiple modalities comes with ready-to-use Docker image user interface. We focus on the stack in its second version, comprised ADAMPro 2.0.0, Cineast interface Vitrivr NG 1.0.0. All components are available under MIT license.
The tremendous increase of multimedia data in recent years has heightened the need for systems that not only allow to search with keywords, but also support content-based retrieval order effectively and efficiently query large collections. In this paper, we introduce ADAM, a system is able store retrieve objects by seamlessly combining aspects from databases information retrieval. ADAM work both structured unstructured jointly provide Boolean similarity search. To handle volumes it makes use...