BAM CT File Format

BAM CT is a file format used at BAM to store either projection or volume data of a CT scan. It is generally used with a two-letter file extensions, with a p as first letter indicating projection data and a b as first letter indicating volume data. The second letter is used to denote the CT scanner the data is captured with.

Header Size and Data Offset

A BAM CT file starts with a 512 bytes long header. This header may be followed by zero-padding until the actual data block begins. The beginning of the raw image data block is located at a multiple of the size of a pixel row (in bytes). If a pixel row needs at least 512 bytes, the data offset can be calculated by

data offset = (number of columns) × (bytes per pixel).

If this data offset is smaller than the significant header size of 512 bytes, the data offset will instead be a multiple of the above-shown row length in bytes:

data offset = N × (number of columns) × (bytes per pixel)

with the smallest integer N that leads to a data offset ≥ 512.

For example, if you have X-ray projections with 1000 pixel columns and 16 bit grey values (i.e. 2 bytes per pixel), the data offset will be at

data offset = 1000 px × 2 bytes/px = 2000 bytes.

Byte Order

Any header or image data unit that is stored in more than 1 byte can be either written in big-endian byte order or in little-endian byte order. The first 12 bytes of the header reveal which byte order is used (see “File Name” section). Since today’s CPUs usually operate in little-endian mode, this is the byte order that BAM CT files generated with aRTist use.

File Name (header entry)

The BAM CT header starts with 12 bytes that contain a “file name” (a string of twelve 1-byte characters, not null-terminated). It follows a structure that reveals additional information:

• character 0 to 6: arbitrary name

• character 7: must be a dot .

• character 8: content type

  • d for projection data

  • b for tomograms (“Bild”)

• character 9: CT device code (a BAM internal ID)

• character 10: image data type:

  • c: 8 bit unsigned int

  • s: 16 bit unsigned int

  • i: 32 bit unsigned int

  • r: 32 bit float (real)

• character 11: byte order:

  • s: little endian

  • x: big endian

Header Data

The following table lists the elements of a BAM CT header.

Note

The element at offset 12 (0xC) represents one of the following:

• for tomograms: the number of rows in an image,

• for projections: (number of rows) × (number of angular steps).

The direction of rotation is given by the sign of the angular step size (offset 176, 0xB0). Its value is positive for counter-clockwise rotations (CCW) and negative for clockwise rotations (CW).

You can calculate the detector’s pixel size from the voxel size (offset 112, 0x70):

pixel size = (voxel size) × SDD / SOD.

Byte	(hex)	Count	Data Type	Description
0	0x00	12	8 bit char	File name, see above
12	0x0C	1	32 bit uint	Number of rows (X) or rows × angular steps, see above
16	0x10	1	32 bit uint	Number of columns (Y)
20	0x14	1	32 bit uint	Number of angular steps
24	0x18	1	32 bit int	Number of angular steps up to 180°
28	0x1C	1	32 bit uint	Number of slices (Z)
32	0x20	1	32 bit uint	Number of translations
36	0x24	1	32 bit uint	Number of intermediate angles
40	0x28	1	32 bit uint	Number of margin points
44	0x2C	1	32 bit uint	Number of detectors
48	0x30	1	32 bit uint	Bytes per pixel
52	0x34	1	32 bit uint	Number of diodes per detector
56	0x38	6	32 bit uint	Reserved
80	0x50	1	32 bit float	Minimum attenuation coefficient [1/cm]
84	0x54	1	32 bit float	Maximum attenuation coefficient [1/cm]
88	0x58	1	32 bit float	Total number of photons
92	0x5C	1	32 bit float	Measurement time per point [s]
96	0x60	1	32 bit float	Velocity number
100	0x64	1	32 bit float	Start angle
104	0x68	1	32 bit float	Scan centre point [mm] (horizontal position)
108	0x6C	1	32 bit float	Scan length without ramp [mm]
112	0x70	1	32 bit float	Voxel size [mm] (sample size)
116	0x74	1	32 bit float	Stage elevation [mm] (vertical position)
120	0x78	1	32 bit float	Elevation increment [mm]
124	0x7C	1	32 bit float	Source-object distance SOD [mm]
128	0x80	1	32 bit float	Source-detector distance SDD [mm]
132	0x84	1	32 bit float	Source elevation [mm] (vertical position)
136	0x88	1	32 bit float	Source centre [mm] (horizontal position)
140	0x8C	1	32 bit float	Source distance [mm] (magnificational position)
144	0x90	1	32 bit float	Detector elevation [mm] (vertical position)
148	0x94	1	32 bit float	Detector centre [mm] (horizontal position)
152	0x98	1	32 bit float	Detector distance [mm] (magnificational position)
156	0x9C	1	32 bit float	Spacer elevation [mm] (thickness)
160	0xA0	1	32 bit float	Object weight [kg]
164	0xA4	1	32 bit float	Beam elevation [mm]
168	0xA8	1	32 bit float	Collimator width [mm]
172	0xAC	1	32 bit float	Collimator height [mm]
176	0xB0	1	32 bit float	Angular step size between images [deg]
180	0xB4	1	32 bit float	PCD clear time per point [s]
184	0xB8	1	32 bit float	Density correction factor [g/cm]
188	0xBC	1	32 bit float	ROI centre [mm]
192	0xC0	1	32 bit float	ROI distance [mm]
196	0xC4	1	32 bit float	Reserved
200	0xC8	8	8 bit char	Source type
208	0xD0	8	8 bit char	Source energy
216	0xD8	8	8 bit char	Source intensity
224	0xE0	8	8 bit char	Detector type
232	0xE8	80	8 bit char	Sample name
312	0x138	4	8 bit char	Program ID
316	0x13C	16	8 bit char	Measurement start time (TT.MM.JJJJ/hh:mm)
332	0x14C	16	8 bit char	Measurement stop time (TT.MM.JJJJ/hh:mm)
348	0x15C	16	8 bit char	Time and date of last edit (TT.MM.JJJJ/hh:mm)
364	0x16C	12	8 bit char	Look Up Table File 1
376	0x178	12	8 bit char	Look Up Table File 2
388	0x184	12	8 bit char	Look Up Table File 3
400	0x190	12	8 bit char	Tube filter
412	0x19C	96	8 bit char	Processing steps
508	0x1FC	4	8 bit char	Reserved
512	0x200			Data or zero-padding until data offset, see above