DICOM Data compression

The Digital Imaging and Communications in Medicine (DICOM) is the most widely used standard for the exchange of medical image data. As a typical study can contain hundreds of files consisting of textual and volumetric data, compression of DICOM files is frequently applied, mostly by using JPEG for the volumetric data within each single file. Here we present Dr. DICOM, an efficient software tool for storing, exchanging and visualizing DICOM files over the Internet. Dr. DICOM consists of a server with stored compressed DICOM studies and a Java applet that performs decompression of volumetric and textual data. The applet offers basic visualization based on ray-casting, selection of colour palettes and basic geometric transformations (zooming and rotating). Three main contributions of Dr. DICOM are as follows.

1. Efficient compression. Files from the same study are considered as a volumetric object, which is progressively compressed using the lossless quadtree-based method from the previous chapter. The average compression rate is around 65%. The text is losslessly compressed using a predicting scheme.

2. Quickly identifiable objects during progressive visualization. Due to the progressively compressed volumetric space, users can quickly determine the content of the data being transferred, progressively decompressed and visualized. Visually useful information is obtained after a few seconds, which is much faster than with classical methods (Zip or Rar), where a complete download is needed before the content can be examined.

3. Interaction with objects during the progresssive decompression. A user can manipulate (rotate, scale) the model while the decompression is still taking place. This is possible as the data is hierarchically transferred, not just as single images.