Dense Image Space is a conceptual, hardware, data and software project, in the early stages of development (2006), designed to acquire and visualize coregistered sets of photographic images from arrays of viewpoints.
Outline
Dense Image Space description
A simple dense image space; acquisition, image, and display geometry
A few possible dense image space configurations
Photographic aquisition basics
Essential aspects of software
coregistration software
display software
possible future software directions
Methods
Photographic hardware
Software platform and components
Image density, memory, and access speed
Comments
DIS masks
So let me see if I'm understanding the implementation of DIS you have proposed
1.) load a series of 80% overlapping images so at some point you will have at least 5 images stacked on top of each other
2.) when the perspective is from the top you just show the top images;
3.) to change the perspective view you use as series of vertical pixel-width masks on the various layers in order to allow the pixel information from other perspective (layers) to be visible on the top level (displayed).
Is this correct? or is it more of a 'when i change perspective, view information from the lower images gets repositioned and displayed on the top image...
Those diagrams a fantastic by the way.
I like seeing this outline too. Its like a teaser and I have to wait for upcoming events... but it helps me see where you are going in terms of a 'bigger picture.' I want to do this with a few of the projects that I’m thinking of but I don’t think I have them as crystallized/formalized as you have DIS. I hope you will eventually also include a section link for example images (screen caps) of the different perspectives and maybe even ultimately some example raw data sets and the DIS viewer so that people could play around with their making their own DIS pictures.
DIS image selection
It's even more simple than the figure might indicate (Fig. 2 got a bit complicated, I haven't decided what to exclude, and I haven't gotten to the caption). There's no mask (3.) involved, more like "when I change perspective, the image that represents the desired perspective is displayed". The figures show ALL perspectives that are available for that bit of subject matter, and the top one is the currently displayed perspective (the one appearing in the viewport).
A few real images mocking up what you might see will help, but will still have to imagine the feel. Simple raw data sets will be easy to construct, but it will be some more work to get a viewer that makes any sense. But a great goal to work toward, even if just for demo purposes.
Post new comment