Closed ant0nsc closed 2 years ago
@peterhessey FYI
Additional details from previous e-mail chain:
"It looks like the problem is near the end of the file.
The dumps show the data twice, first as hex, then as ASCII. Each pair of lines is prefixed with the byte offset.
After the end of the file, the original file has extra 00 bytes, as shown in the snippets below. These were cut and pasted from the hex dump files and had the byte offset corrected.
If you want to test it, you can use DICOM+ to read it in and see if it shows any errors.
Original file:
1232651 65 74 65 72 5e fe ff 0d e0 00 00 00 00 fe ff dd e0 00 00 00 00 **00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00**
e t e r ^ ~ del cr ` nul nul nul nul ~ del ] ` nul nul nul nul **nul nul nul nul nul nul nul nul nul nul nul nul nul nul nul nul**
Fixed file:
1232677 65 74 65 72 5e fe ff 0d e0 00 00 00 00 fe ff dd e0 00 00 00 00
e t e r ^ ~ del cr ` nul nul nul nul ~ del ] ` nul nul nul nul
"
It was recommended that the data between the pairs of asterisks (**) should be removed (asterisks added manually).
Adding @furtheraway for visibility, who had originally posted the question
Discussed in https://github.com/microsoft/InnerEye-DeepLearning/discussions/687
AB#5453