Itk Data Source

Http://blog.csdn.net/zhangcunli/article/details/4678711

Data sources for Medical Image research:

1. rire:
J. m. professor fitzpatriek leads the repeatable Medical Image Registration evaluation (rire) project team, which is the world's most well-known and complete medical image registration evaluation. the project provides a large amount of patient instance data, mainly used for CT-MR and PET-MR registration evaluation. researchers can download their data, use their own registration technology for the experiment, submit their experiment results to the website, after evaluation by professional staff, the evaluation results will be returned to the investigator.

Lab data:
Http://www.insight-journal.org/RIRE/download_data.php
The downloaded data is not in DICOM format. To use DICOM format data, you can use a tool provided by the website to convert it to DICOM format. The format conversion example is as follows:
Conversion format:> imagereaddicomserieswritewin32.exe sourceimage outdir
For example, we downloaded the data of patient patient_109:
Decompress ct1_1).tar.gz in patient_109. If image. Bin is not decompressed, decompress image. bin. Z.
Replace sourceimage with patient_109_ct.mhd to convert the data into DICOM files. outdir is the output directory of DICOM files.

2. 2d-3d registration standard evaluation, address: http://www.isi.uu.nl/Research/Databases/GS/
I have not downloaded the data provided by this website.

3. Data Links officially provided by Itk
Http://public.kitware.com/pub/itk/Data/
They are:
Brainweb, simulated brain database: http://public.kitware.com/pub/itk/Data/BrainWeb/
Data for Finite Element Method experiments: http://public.kitware.com/pub/itk/Data/DeformableRegistrationFEM/
Liver Tumor: http://public.kitware.com/pub/itk/Data/LiverTumor/
Registerimagemodulesbaselines:
Http://public.kitware.com/pub/itk/Data/RegisterImageModulesBaselines/
Visiblewomanhead:
Http://public.kitware.com/pub/itk/Data/VisibleWomanHead/
And so on. one problem is that most of the data is. MHD ,. the raw format does not show how the data is processed after half a day. Raw should be the raw data, and the MHD specifies how to read the data in the raw file. after trying, you can use VTK to visualize this type of data. However, raw files generally contain a large amount of data and there is no effective solution yet. however, this type of data is used in the examples of deformation registration in Itk. I do not know where to introduce this format.

4. DICOM Dataset: http://pubimage.hcuge.ch: 8080/
This website has a large volume of data, which varies with different locations and conditions.

5. Stanford University: http://graphics.stanford.edu/data/voldata/ is strange,
The downloaded data is clearly associated with the DICOM inspection software, but cannot be opened, and a lot of software cannot be opened, depressing.

6. Visual Human Project CT Database: https://mri.radiology.uiowa.edu//VHDicom/, which provides two different formats of data,
The following link is in DICOM format.

7. brainweb: The simulated brain database simulates the brain database. The above official Itk link also provides a brainweb, but I don't know if it is the same. Some parameters need to be selected for downloading data provided by the website, I still don't know what it is. this website is very famous for brain image data, but I still don't know how to handle the data formats it provides. Alas, there is a lack of knowledge and materials in this area...
Http://www.bic.mni.mcgill.ca/brainweb/

8. Institute of automation, Chinese Emy of sciences:
Http://www.mitk.net/zhcn/index.htm, one of the data is dedicated to their 3 dmed,
The other database is micro-CT mouse data. The two links are bad, or I sent them an email a few days ago before they fixed the problem. It's really irresponsible.

This is basically the case. After reading some papers, most of Medical Image Registration uses rire projects, and some use brainweb brain image data. what is important is that rire will judge the investigator's experiment results, and the data has been tested by a large number of experts and should be the most authoritative.

The trouble is that medical information is too lacking, and many data formats do not understand what is going on. After a long time, I found that I did not find out how to deal with it. this will be put down for the moment, and I will talk about it later.