April 20, 2007
README
Disclaimer
The test data are provided “as is” to users as a courtesy. In no event will the CCSDS, its member Agencies, or the provider of the data be liable for any consequential, incidental, or indirect damages arising out of the use of or inability to use the data.
Description
The provided data is used to test implementation of CCSDS 122.0-B.1 in different configurations. Effort was made to cover most cases but the coverage is not exhaustive, particularly in the extreme values of parameters such as ImageWidth, segment size S, SegByteLimit, etc.
For implementations that produce output as single-byte words, download the folder OutputByte (the zip-compressed version).
For implementations that produce output as 2-byte words, download the folder Output2Byte (the zip-compressed version).
Table 1 lists the thirteen sub-directories within each folder. It is suggested that testing start from sub-directory test_hdr1_xxxx, then proceed to test_hdr234_xxxx and other folders.
The data were created using the “big-endian” convention for storing data. Users are cautioned that data byte swapping may be necessary when using these data on different platforms. However, the “endian” convention does not change the bit sequence in the coded files.
For some test images, the DWT coefficient files are provided as well. These are either stored as integer*4 for integer DWT, or float*4 for float DWT. These DWT files have file extension “.block.coef” which contains localized 8x8 coefficient blocks arranged in raster order. To display the DWT coefficients in these files, they have to be read into display program line by line with the same ImageWidth (or appended ImageWidth so it is multiples of 8) as for the test image. This format is used to facilitate visualization and debugging, and does not correspond to the sequence in which coefficients are processed by the Bit-Plane-Encoder specified in Chapter 4 of CCSDS 122.0-B.1.
For lossy compression, users may choose to implement DWT reconstruction scheme different from what is suggested in CCSDS 120.1-G.1, in which case users may obtain measurements of distortion on the reconstructed images different from the values provided in the test sets.
Table 1: Summary of Test Data
Test Number / Sub-directory Name / Test Description1 / test_hdr1_byte/test_hdr1_2byte / Lossless on smaller and full-size images using only Part 1 header
2 / test_hdr234_byte/test_hdr234_2byte / Lossless test, using Part 2, 3 and 4 headers
3 / test_hdr234_1bTO4b_lossless_byte/
test_hdr234_1bTO4b_lossless_2byte / Lossless test on images with 1-bit, 2-bit, 3-bit and 4-bit input range, using Part 2, 3 and 4 headers
4 / test_hdr234_433_byte/test_hdr234_433_2byte / Testing section 4.3.3. Lossless mode using Part 2, 3 and 4 headers
5 / test_hdr234_ExtImage_byte/test_hdr234_ExtImage_2byte / Image with random data or all-zero data. Lossless using Part 2, 3 and 4 headers
6 / test_hdr234_MaxDynamicRange_byte/
test_hdr234_MaxDynamicRange_2byte / Lossless test on image which produces maximum dynamic range after integer DWT. Part 2, 3 and 4 headers used.
7 / test_hdr234_lossy_byte/test_hdr234_lossy_2byte / Lossy test at different coding rates using integer DWT. Part 2, 3, and 4 headers used.
8 / test_hdr234_mode_byte/test_hdr234_mode_2byte / Lossy test at different quality levels using integer DWT. Part 2, 3 and 4 headers used.
9 / test_hdr234_float_byte/test_hdr234_float_2byte / Lossy test using float DWT. Part 2, 3 and 4 headers used.
10 / test_hdr234_MaxDynamicRange_floatDWT_byte/
test_hdr234_MaxDynamicRange_floatDWT_2byte / Lossy test on image which produces maximum dynamic range after float DWT. Part 2, 3 and 4 headers used.
11 / test_hdr234_CustW_byte/test_hdr234_CustW_2byte / Lossless test with user supplied weights for integer DWT. Part 2, 3 and 4 headers used.
12 / test_hdr234_Uniform_byte/test_hdr234_Uniform_2byte / Lossless test on a 16-bit uniform intensity image
13 / test_hdr234_oddsize_byte/test_hdr234_oddsize_2byte / Lossless test on an image with odd-size
1