In contrast with MIAS database where a single scanner was used to
digitize the mammograms, in DDSM four different scanners were
used. Thus, depending to which of them was used, the mammograms
have different characteristics. Table
shows some of these parameters, like the number of bits of the
image and the sampling rate.
Each case of the database is named as M_XXXX_
, where ``M''
represents the used scanner (from ``A'' to ``D''), XXXX is a
number of four digits. The four images of the case are stored in a
Lossless JPEG format, and the own database provides software to
transform the images in a typical raw format.
The annotations stored in each case includes information related to the image (date of study, kind of film), related to the patient (patient's age, density of the breast in BIRADS), related to each image of the case (size), and an overlay file if necessary. In such file, the number of abnormalities and their description (kind of abnormalities, assessment, boundary) are explained.