DimP - A Direct Manipulation Video Player


DimP is a video player prototype that allows to browse video clips by directly manipulating their content.

The way it works is quite simple:

  1. DimP automatically extracts object motions from the videos.
  2. DimP then allows the user to control video playback by scrubbing these objects on their trajectories.

Videos

One-minute teaser video.
[youtube]   [download]
Longer video with explanations.
[youtube]   [download]

Demo

You can download DimP for free (Windows). Mac OS users, check this website.

Paper

Dragicevic, P., Ramos, G., Bibliowitcz, J., Nowrouzezahrai, D., Balakrishnan, R., and Singh, K. 2008. Video Browsing by Direct Manipulation. In Proceedings of the Twenty-Sixth Annual SIGCHI Conference on Human Factors in Computing Systems (Florence, Italy, April 05 - 10, 2008). CHI '08. ACM, New York, NY, 237-246.
[BibTex and DOI here].

You can also download the slides from our talk at the CHI '08 conference (60MB Zip file with pptx, pdf and avi files). You are free to reuse this content for your own purposes (license here).

Is this Really Direct Manipulation?

Yes. Although the video clip itself is not modified, dragging an object constrained to a trajectory is similar to dragging a slider. The slider's thumb will never stray away from its predefined trajectory, yet we all agree it is being directly manipulated.

Whether it is true direct manipulation or not is actually an ill-posed question. On computer screens, where everything is just pixels, you are never physically manipulating objects. Direct manipulation is always an illusion produced by having user's gestures match the resulting motions on the screen as closely as possible. This is further discussed in our paper above.

So what's being manipulated, exactly? Both the video content (i.e., the things you see moving in the video) and the "tape head". When using DimP, the user directly manipulates the video content and indirectly manipulates the tape head. When using the seeker bar, the user directly manipulates the tape head and indirectly manipulates the video content.

Related Videos

The idea of manipulating autonomous entities defies the ordinary and as such, has been explored by a few artists. Here are three nice video clips which play with this paradox.

Daniel Chesterfield by Chris Van den Durpel.
[youtube]
Pro X Fade Ad by Eclectic Breaks.
[youtube]
I Parking by June Bum Park.
[youtube]

Similar and Parallel Work

The idea of using direct manipulation for controlling video playback was first introduced in 1999 by Takashi Satou and his colleagues from the Japanese company NTT-AT [1,2,3]. Their system required video motion to be specified manually but other than that, the idea was already there.

Takashi Satou's work probably did not get the attention it deserved from the HCI community. Eight years later, four research groups, including ours, rediscovered the idea in parallel. We were all working on a fully automatic support for video browsing by direct manipulation, with technically different approaches but also striking similarities in the thought process.

While our group was still experimenting with a Java prototype, Dan Goldman and his colleagues from the University of Washington presented a video storyboarding system at SIGGRAPH' 06 [4], which features an "arrow scrubbing" technique for browsing videos. Although scrubbing occurred in a separate window, they later implemented a direct manipulation variant they describe in a technical report [5].

At about the same time, our DimP system was ready and we submitted a paper to UIST '07, but it has been unfortunately rejected. A few months later, Don Kimber and his colleagues from FXPAL presented a short paper at ICME '07 about a video surveillance playback system with support for video browsing by direct manipulation [6].

We finally presented our work at CHI '08 [7]. In the same session, Thorsten Karrer and his colleagues from the RWTH Aachen University presented a short paper about another similar system called DRAGON [8]. See below for their subsequent work.

[1] Satou, T., Kojima, H., Akutsu, A., and Tonomura, Y. (1999). CyberCoaster: Polygonal line shaped slider interface to spatio-temporal media. ACM MULTIMEDIA '99. [www]
[2] NTT-AT. (2002). Dragri. [www]
[3] Satou, T., Kojima, H., Tonomura, Y. Akutsu, A. (2003) Scheme for graphical user interface using polygonal-shaped slider. United States Patent 6542171.
[4] Goldman, D.B., Curless, B., Salesin, D., and Seitz, S.M. (2006). Schematic storyboarding for video visualization and editing. SIGGRAPH '06. [www]
[5] Goldman, D.B., Curless, B., Salesin D., Seitz, S.M. (2007). Interactive Video Object Annotation. Tech Report UW-CSE-2007-04-01. [www]
[6] Kimber, D., Dunnigan, A., Girgensohn, A., Shipman, F., Turner, T., TaoYang. (2007). Trailblazing: Video Playback Control by Direct Object Manipulation. ICME '07.
[7] Dragicevic, P., Ramos, G., Bibliowitcz, J., Nowrouzezahrai, D., Balakrishnan, R., and Singh, K. (2008). Video Browsing by Direct Manipulation. CHI '08.
[8] Karrer, T., M. Weiss, Lee, E., Borchers, J. (2008). DRAGON: A Direct Manipulation Interface for Frame-Accurate In-Scene Video Navigation. CHI '08. [www]

Subsequent Work

20 Oct 2008: Dan Goldman (now at Adobe's Advanced Technology Labs) presented a fully-developed version of his system at UIST '08 [1]. It supports, among other things, robust object tracking and multi-point motion constraints.
30 Oct 2008: Moritz Wittenhagen from the RWTH Aachen University submitted his Master's thesis on DragonEye [2], a system for video browsing by direct manipulation that supports fast object tracking, handles occlusions, and compensates for camera motion.
25 Nov 2008: Dan Goldman posted a video demo of his system on Vimeo.
9 Fev 2009: Christian Brockly from the RWTH Aachen University submitted his Master's thesis on evaluating direct manipulation techniques for video navigation [3].
Sept. 2009: Throsten Karrer and his colleagues from the RWTH Aachen University presented a mobile version of DRAGON [4].
Nov. 2010: Myunghee Lee and Gerard J. Kim from POSTECH in Korea presented Empatheater, a system that might well be a precursor of whole-body direct manipulation video players (video here) [5].
[1] Goldman, D.B., Gonterman, C., Curless, B., Salesin, D. and Seitz, S.M. (2008). Video Object Annotation, Navigation, and Composition. UIST '08. [www]
[2] Wittenhagen, M. (2008). DragonEye - Fast Object Tracking and Camera Motion Estimation. RWTH Aachen University Master's thesis. [www]
[3] Brockly, C. (2009). Evaluation of direct manipulation techniques for in-scene video navigation. RWTH Aachen University Master's thesis.
[4] Karrer, T., Wittenhagen, M., and Borchers, J. (2009). PocketDRAGON: a direct manipulation video navigation interface for mobile devices. MobileHCI '09.
[5] Lee, M. and Kim, G. J. (2010). Empathetic video experience through timely multimodal interaction. ICMI-MLMI '10. [www]

Contact

  • Pierre Dragicevic
  • Gonzalo Ramos

    Please post your questions and comments to our discussion group.