Omar Shahbaz Khan

Hi!

Thanks for dropping by my site.

I am a computer science researcher specializing in large-scale multimedia retrieval and analytics. My work focuses on how people search and interact with massive multimedia collections, and on building retrieval systems that balance performance, scalability, and efficiency.

I develop multimedia systems that combine foundation models, high-dimensional indexing, and databases to support both exploration (e.g., browsing and summarizing) and search (e.g., query-based retrieval).

A key theme in my research is resource-aware system design. That means understanding and optimizing how storage, memory, and compute demands evolve as collections grow, so systems scale efficiently without simply adding more hardware.

I am also interested in understanding human information-seeking behavior and how it might inspire new forms of adaptive or strategy-driven retrieval systems.

You can reach me at omsh@itu.dk, or at osk@oskhan.com for anything outside work.

News

  1. Oct 2026

    Exquisitor won Video Retrieval for Beginners (VR4B) at CBMI 2026, a competition where audience members, not expert users, operate the systems.

  2. Jun 2026

    Two papers at ICMR 2026 in Amsterdam, one on what drove success at VBS 2026 and one on optimizing long-running media aggregation queries.

  3. Jan 2026

    Exquisitor took 2nd place at the Video Browser Showdown (VBS) 2026 in Prague. The system paper, Temporal Queries Revisited, is available in the MMM 2026 proceedings.

  4. Nov 2025

    Started as Assistant Professor at the IT University of Copenhagen.

  5. Oct 2025

    multiXview won the CASTLE 2025 Interactive Grand Challenge at ACM Multimedia 2025 in Dublin.

  6. Jun 2025

    Exquisitor won the interactive track of the IViSE workshop at CVPR 2025.

Awards