DCU has strength and depth in research in both video and still digital imaging technologies. At the core is the Centre for Digital Video Processing (CDVP) although the recently approved Clarity research initiative has expanded the focus area to include sensor networks.

The CDVP mission is to research and develop techniques and tools to automatically analyse and index digital video information and allow content-based operations such as browsing, searching, alerting, filtering and summarisation.

Multimedia Research at DCU

MediAssist Photo-management
MediAssist - Software tool to enable easy and automatic management of digital photo collections based on content (people, buildings, faces) and context information (location, time, date).  Easy-to-use intuitive web-based interface.

Adaptive Object Detection and Tracking
Adaptive object detection and tracking software solution that uses multiple inputs (e.g. visual, thermal, audio) for improved tracking, for security and automated surveillance applications.

Pedestrian Tracking
Pedestrian detection and tracking using stereo vision (two cameras). Solution robustly handles variability in lighting conditions, occlusions and variation in camera positions. Applications are likely to be security related and also for assessment of foot traffic in public places e.g. airport, public transport and transit points, shopping malls.

Image Matching
Image matching software that robustly and quickly matches a digital image of an object against a database of images. Example applications to date include - word matching in scanned hand-written documents, searching CAD & blueprints for specific features. Many others possible.

Technologies for Efficient Access to Multimedia Content from Mobile Phones
Technologies and know-how, to access archives of photos and video on a mobile, overcoming handset limitations of screen size, processing power and interface. Complete technology solution, leveraging core expertise in search engine design, visual processing tools to better understand multimedia content, user interface design and hardware development.

Search Technology for Video Archives Using Pictures and Speech as Queries
Search technologies enable a picture, speech or free text (or all three) to be used to query a video archive. Database of content is sorted automatically using content analysis algorithms that automatically extract useful information from video that can then be searched against. Avoids need for manual tagging of multimedia clips, the main approach to date.

Tools for Synchronous Collaborative Web Search
Collaborative search algorithms to allow mutliple users, independent of location, to carry out shared searches more effectively than as individuals.

Life Logging
A software solution that effectively analyses, organizes and summarizes huge amounts of digital photographs (or video clips) into meaningful visual summaries. The best known implementation is using the SenseCam that a user wears all the time and which automatically takes photographs passively approx every 20 secs. More sophisticated approaches can use video, and can integrate other data - GPS, interactions with others via bluetooth, and incorporation of other Internet content, to enable even more compelling applications.

Search Engine Optimization Techniques
Algorithms to enable improved search engine performance through usage of document parameters to indicate relevance and quality. These are then used to eliminate large numbers of low quality documents returned in a typical search, prioritize most relevant documents, hence improving overall search performance.

Gesture Recognition for Mobile Phones
Gesture recognition software as interface to mobile phones and gaming consoles. 3D gesture recognition has advantages over traditional keypad and touch screen inputs.

Back to top