From CollectiveAccess Documentation
Jump to: navigation, search

Icons-flag-fr.png TilePic_FR

The TilePic image file format

From the TilePic web site at the Digital Library Project, UC Berkeley:

TilePic is a file format developed by the Berkeley Digital Library Project that is designed to store tiled data of arbitrary type 
in a hierarchical, indexed format in order to provide fast retrieval. This format was influenced by the Kodak FlashPix format 
and is based in part on the GridPix format developed by the Tertiary Disk Project at UC Berkeley.

While indeed able to store any type of data, TilePic is most often used to store images as sets of fragments (or "tiles") at multiple resolutions. A TilePic-compatible image viewer can implement efficient pan-and-zoom display images of virtually any size by only requesting the tiles needed for display at the minimum resolution required for quality viewing.

CollectiveAccess can automatically convert uploaded images into TilePic files containing JPEG'ed tiles in a series of resolutions up to and including the resolution of the original uploaded image. CollectiveAccess's built-in image viewer implements a lightweight easy-to-use TilePic compatible viewer with continuous zoom and annotation capabilities.

You can learn more about the TilePic format at http://elib.cs.berkeley.edu/tilepic/

Also, a few things about Tilepic encoding in CollectiveAccess:

  1. Very large images can use a whole lot of memory while converting. Generally ImageMagick will try to grab about 3 times the *uncompressed* size of the image while it's working. For a 200meg jPEG image, about 4.2gigs is grabbed by ImageMagick. ImageMagick should be able to process with less memory than the required memory if need be, but will generally take longer to complete if that is the case.
  2. Tilepic conversion is supported when running GD instead of ImageMagick, but that hasn't been tested extensively. From we've heard GD doesn't tolerate really large images very well. If you're planning to use CollectiveAccess with large images definitely stick with ImageMagick or CoreImage as the image processing back-end if at all possible.

(See also the open source TilePic reader for Windows.)


Personal tools