Wiki
Clone wikiuocte / Heidelberg File Format
Because no specification of this file format was available for the development of uocte, the file format was reverse engineered for interoperability. The information on this page therefore is incomplete and may be incorrect. It only serves to document which parts of the data are interpreted by uocte and which assumptions it makes concerning interpretation.
Heidelberg data is stored in a single binary, little endian file with extension e2e or E2E. It contains a header, a directory that is split in chunks of entries in a single-linked list, and data chunks. The high-level structure is this:
// header char[12] magic1 u32 version ("0x64") u16[9] 9x "0xffff" u16 "0x0" // main directory char[12] magic2 u32 version ("0x64") u16[9] 9x "0xffff" u16 "0x0" u32 num_entries u32 current u32 "0x0" u32 ? // traverse list of directory chunks do seek(current) char[12] magic3 u32 version ("0x64") u16[9] 9x "0xffff" u16 "0x0" u32 num_entries u32 ? u32 prev u32 ? for i in num_entries..1 u32 pos u32 start u32 size u32 "0x0" u32 patient_id u32 study_id u32 series_id u32 slice_id u16 ? u16 "0x0" u32 type u32 ? if start > pos push (start,size) to stack S current = prev while current != 0
This gives a stack of data chunks to process. uocte uses this first pass through the file to determine the number of slices as the maximum of slice_id/2 for each combination of study_id and slice_id. uocte assumes only one patient_id to occur in a file.
In a second pass, uocte seeks to each the position from the stack S and reads one data chunk. Each chunk has a header with the following structure:
u8[12] magic4 u32 ? u32 ? u32 pos u32 size u32 "0x0" u32 patient_id u32 study_id u32 series_id u32 slice_id u16 ind u16 ? u32 type u32 ?
slice_id is 0xffffffff for data not specific to a slice. There are numerous data chunks of differnt type, most of which I could not determine their semantics. Only the tags processed by uocte are listed here in
detail. All strings are treated as ISO8859-1 encoded.
// type == 0x00000009 - patient info u8[31] given name u8[66] surname u32 birthdate u8 sex ("M" or "F")
The date of birth as Julian date is: JD = (birthdate/64) - 14558805.
// type == 0x0000000B - laterality u8[14] ? u8 laterality ('L' or 'R') u8[?] ?
// type == 0x40000000 - image data u32 size u32 type ("0x02010201" for fundus, "0x02200201" for tomogram) u32 ? u32 width u32 height if ind == 0 u8[height][width] raw fundus image else uf16[height][width] raw tomogram slice image
uf16 is a floating point type with no sign, 6-bit exponent, and 10-bit mantissa. uocte assumes that the tomogram spans [1/6,5/6]x[1/4,3/4] of the fundus image and has x- and z-dimensions of 6 and 4.5 millimeters, respectively, an a y-resolution of 3.9µm. These values were inferred from exported XML files. The images are not yet correctly registered. The official viewer shears the images in y-direction. The amount of shear has not yet been found in the data.
// type == 0x00002723 - contour data u32 ? u32 id u32 ? u32 width f32[width] contour depth in tomogram pixels for current slice
Note that the z-direction for tomogram and contours needs to be flipped.
Not yet found in the data is the aquisition date.
Updated