Andreas Girgensohn's Research History

Home Page


In 1996, I joined FX Palo Alto Laboratory. Here is a history of some of my older research projects.

StainedGlass Collage. Built a Web site for creating collages reminiscent of stained glass windows from uploaded images.

Digital Photo Organizer. Created a novel application for organizing large numbers of digital photos. The research challenges were to identify ways for users to group the photos into meaningful categories and to design an intuitive and effective user interface. The application exploits time stamps, image content, image analysis attributes to support users in browsing, filtering, organizing thousands of photos.

Video Editor. Designed and developed a video editing system, called Hitchcock, that supports users in creating custom videos from raw video shots made with a standard DV camera. A recent version supports the creation of detail-on-demand videos with hyperlinks between video clips. This format is also used for automatically generated multi-level video summaries. The research challenge was to make video editing easy by balancing task automation with user control. I developed a semi-automatic solution that includes automatic video analysis to find and trim the best quality video clips, an algorithm to cluster those clips into meaningful piles, and an intuitive user interface for combining clips into a final video and for creating detail-on-demand videos. Key innovations of the approach include video analysis techniques for camera motion estimation, video segmentation and extraction of the most suitable video portions, a user interface presenting video clips in piles, the concept of detail-on-demand videos, and the automatic generation of multi-level hypervideo summaries. User studies and user feedback informed the evolution of the system.

Video Access System. Explored and developed a Web-based digital library, called MBase, for browsing and managing collections of digital videos. The research challenges were to create video summarization techniques and to combine them with flexible user interaction mechanisms to enable users to explore large video collections. I developed automatic algorithms to identify such information as speakers and shot boundaries, to select key video images from a video, and to present that information in a form suitable for browsing. I also created a comic-book-like visualization that summarizes videos and provides entry points into them. The main innovation of this work was the use of keyframes, information extraction, video summarization for efficiently browsing recorded video content, as well as an efficient layout algorithm for the visual summaries. The MBase technologies were used for the CSCW 2000 and Multimedia 2002 / 2003 video programs and are the basis for products being developed by Fuji Xerox (MediaDEPO).

Social Web Sites. Created Web-based tools and sites that foster social interaction, support collaboration, facilitate awareness. The research challenges were to explore the nature of and to develop user mechanisms for Web-based, human-human interaction. CHIplace is a conference Web site built with technologies such as JavaServer Pages and JDBC that facilitates social interaction, social browsing, social networking. It was designed using a system framework that separates the social interaction components from the functional elements and evolved based on analysis of usage logs and user feedback. Its codebase has been open-sourced and was used by CSCW 2002 for CSCWplace.

Reauthoring of Web Pages for Mobile Devices. Created a system for reauthoring Web pages such that they can be displayed efficiently on small, mobile devices such as PDAs and cell phones. The challenge was to develop a framework that could correctly handle a wide variety of Web pages. This system was considered for inclusion into the product line-up of Unwired Planet (now Openwave).



nynex logo

From 1992 to 1996, I worked at NYNEX Science & Technology (now part of Verizon) in White Plains, NY.


Portholes. NYNEX Portholes is a video-based background awareness application created to support groups of people in different physical locations. The goal of our version of Portholes was to explore the benefits of video in supporting group awareness and communication and to make Portholes generally accessible so that it can be studied in a real-world setting.

Dynamic Forms. Designed and created a representation language and user interaction abstraction for complex, highly-interactive, dynamic and Web-based forms. The key features are dynamic field visibility and layout based on user input, early validation of field values, computation of field values, and repetition of fields. The dynamic forms abstraction is a significant HCI contribution that is seamlessly integrated into the usage model and technology framework of the Web.

System for Business Marketing Representatives. Led a technical effort at NYNEX to implement a new user interface for 8000 customer service representatives taking phone service orders. This effort used a participatory design process that revealed many insights into the needs of users with a complex data acquisition task and who also need to interact with live customers and mainframe databases.

Expectation Agents. Created Expectation Agents that enabled developers to formulate assumptions about the use of the system that were verified by software agents observing the user interactions. Such agents overcome limits with user participation in which a small developer team can only involve a limited number of users.



cu logo

Before I joined NYNEX, I received a Ph.D. degree in computer science from the University of Colorado at Boulder (1988 – 1992). My Ph.D. research with Gerhard Fischer in the Human Computer Communication Research Group focused on supporting end-user modifiability in knowledge-based design environments. End-user modifiability allows different users to tailor a system in order to pursue different tasks, to provide different preferences, and to adapt a system to a user’s changing needs.