28.02.2013 Views

Bio-medical Ontologies Maintenance and Change Management

Bio-medical Ontologies Maintenance and Change Management

Bio-medical Ontologies Maintenance and Change Management

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

2.3 Extracting Data from DICOM File<br />

Multimedia Medical Databases 83<br />

In order to extract data from the DICOM file, we are taking into account every tag<br />

from the DICOM dictionary. The tag will be searched in the file <strong>and</strong> in case of<br />

finding it the corresponding value will be extracted.<br />

The steps of extracting information from DICOM files, designed at Software<br />

Engineering Department, Faculty of Automation, Computers <strong>and</strong> Electronics,<br />

Craiova, are:<br />

1. Verifying the existence of the ‘D’,’I’,’C’,’M’ characters in the file prefix.<br />

2. Establishing the type of VR (ExplicitVR or ImplicitVR). This information is<br />

given by the UID (Unique Identifier), information stored in the value field corresponding<br />

to the Transfer Syntax Tag.<br />

3. Establishing the Byte Ordering (BigEndian or LittleEndian). The information<br />

is also given by UID, stored in the value field of the same Transfer Syntax Tag.<br />

The DICOM st<strong>and</strong>ard contains all the values that UID can have.<br />

4. Searching a tag in DICOM file according to the VR type <strong>and</strong> ByteOrdering<br />

5. Value extraction of the corresponding found tag.<br />

DICOM St<strong>and</strong>ard contains over 20 types of binary data or ASCII. The type of<br />

information stored in the value field is given by VR. In accordance with this type<br />

will be extracted strings, integer or byte type information.<br />

Next it is described the problem of image extracting from the st<strong>and</strong>ard DICOM<br />

files, taking into account the method of compression that was used: RLE, JPEG. A<br />

single DICOM object can contain only one image, but that image may have<br />

multiple “frames”, allowing storage of cine-loops or other multi-frame data.<br />

The images from DICOM files can be classified using several criteria:<br />

1. The number of images stored in a file: single frame or multi-frame.<br />

2. Number of bits per pixel: 8 bits, 12 bits, 16 bits or 24 bits.<br />

3. Compression: without compression (raw) or with compression (RLE JPEG,<br />

JPEG Lossless, JPEG 2000. LZW (zip) compression can be used for the whole<br />

data set (not just the pixel data) but is rarely implemented.<br />

4. Photometric interpretation: gray scale images, color images, palette color<br />

images.<br />

In the images without compression, the extraction of pictures is made pixel-bypixel,<br />

taking into account the number of bits stored for each pixel <strong>and</strong> the<br />

photometric interpretation (for monochrome images a pixel is stored using<br />

maximum 2 bytes <strong>and</strong> for color images, a pixel is stored using 3 bytes). In the<br />

images that use compression it is necessary a decompression algorithm before<br />

saving.<br />

The pseudo code for retrieving the frames is:<br />

Set variable number to 0<br />

Loop until all frames are retrieved<br />

Set file dimension as Rows*Columns*SamplePerPixel<br />

Read all file dimension pixels from file starting<br />

With (Header Length + number* file dimension)

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!