New Mammography Dataset - INBreast

[Inês Domingues]

Hello everyone,

Why dont you start using the INBreast database?

http://www.sciencedirect.com/science/article/pii/S107663321100451X

Best regards,


Inês Domingues
-----------------------------------------------------------------------------------------------
PhD student
INESC TEC and Faculdade de Engenharia, Universidade do Porto
https://sites.google.com/site/indezitindezit/
-----------------------------------------------------------------------------------------------

On Wed, Sep 26, 2012 at 6:58 AM, Mehul Sampat <mehul....@ieee.org> wrote:
> Safi,
> LJPEG is a compressed file (using lossless JPEG compression).
> http://marathon.csee.usf.edu/Mammography/DDSM/case_description.html#LJPEGFILE
>
> you need to uncompress it first before trying to load it with the code
> snippet from my previous email.
> an old version of the un-compression code is available here:
> ftp://figment.csee.usf.edu/pub/DDSM/software/bin/jpeg.
>
> if the compression works it will create a file called
> C_0029_1.LEFT_CC.LJPEG.1
> let us know if this works for you.
> good luck,
> Mehul
>
>
> On Tue, Sep 25, 2012 at 7:08 PM, SAFI ULLAH Marwat <safi...@gmail.com>
> wrote:
>>
>> Hello guys,
>>
>> I have the same problem in extracting the image files. I found two things
>> while using the matlab script provided in the thread:
>> No.1. The read image have less dimentiones than the given ones. for
>> example for the case benign_01\case0029\C_0029_1.LEFT_CC.LJPEG the ics file
>> reads the lines (rows) as 4648 while columns as 2672 but when I read the
>> file it shows the size of the image as 4648x1772
>>
>> No.2. the decompressed image carries no information, i mean it is just a
>> collection of random texturel dots.
>>
>> any help will be highly counted on.
>>
>> Regards,
>>
>> Safi
>>
>>
>>
>> On Tue, Jan 3, 2012 at 2:08 PM, Praful Agrawal <pra...@iiitd.ac.in>
>> wrote:
>>>
>>> It worked..thanks a lot
>>>
>>> On Jan 3, 1:56 pm, Mehul Sampat <mehul.sam...@ieee.org> wrote:
>>> > ps: oops the first line in the code snippet should be: fid =
>>> > fopen(filename,'r','ieee-be');
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> > On Tue, Jan 3, 2012 at 12:47 AM, Mehul Sampat <mehul.sam...@ieee.org>
>>> > wrote:
>>> > > Hi Praful,
>>> >
>>> > > The uncompressed images should not have a lot of noise. Most likely
>>> > > it
>>> > > could be
>>> > > a big endian vs. little endian issue. (I believe you need to read it
>>> > > in as
>>> > > big-endian)
>>> > > here is a snippet of matlab code to read the *.1 file. (you can
>>> > > change it
>>> > > to c if needed)
>>> > > -------------------------------------------
>>> > > fid = fopen(fid,'r','ieee-be');
>>> > > image = fread(fid,[COLS ROWS],'short');
>>> > > image = image';
>>> > > image = uint16(image);
>>> > > -------------------------------------------
>>> > > where COLS,ROWS is the number of columns/rows in the image; get this
>>> > > info
>>> > > from *.ics file.
>>> > > Please let the group know if this works for you.
>>> > > Good luck
>>> > > Mehul
>>> >
>>> > > On Tue, Jan 3, 2012 at 12:26 AM, Praful Agrawal
>>> > > <praf...@iiitd.ac.in>wrote:
>>> >
>>> > >> Hi all,
>>> >
>>> > >> I have downloaded DDSM database, but the decompressed images have a
>>> > >> lot of noise in it. I am new to this area. Please help me.
>>> >
>>> > >> Thanks and Regards,
>>> > >> Praful
>>> >
>>> > >> --
>>> > >> You received this message because you are subscribed to the Google
>>> > >> Groups
>>> > >> "Mammographic Image Analysis Research Community" group.
>>> > >> To post to this group, send email to mammo...@googlegroups.com.
>>> > >> To unsubscribe from this group, send email to
>>> > >> mammoimage+...@googlegroups.com.
>>> > >> For more options, visit this group at
>>> > >>http://groups.google.com/group/mammoimage?hl=en.
>>>
>>> --
>>> You received this message because you are subscribed to the Google Groups
>>> "Mammographic Image Analysis Research Community" group.
>>> To post to this group, send email to mammo...@googlegroups.com.
>>> To unsubscribe from this group, send email to
>>> mammoimage+...@googlegroups.com.
>>> For more options, visit this group at
>>> http://groups.google.com/group/mammoimage?hl=en.
>>>
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "Mammographic Image Analysis Research Community" group.
>> To post to this group, send email to mammo...@googlegroups.com.
>> To unsubscribe from this group, send email to
>> mammoimage+...@googlegroups.com.
>> For more options, visit this group at
>> http://groups.google.com/group/mammoimage?hl=en.
>
>
> --
> You received this message because you are subscribed to the Google Groups
> "Mammographic Image Analysis Research Community" group.
> To post to this group, send email to mammo...@googlegroups.com.
> To unsubscribe from this group, send email to
> mammoimage+...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/mammoimage?hl=en.

[SAFI ULLAH Marwat]

Dear Ines,

 

Can you please give us a lilnk toward the dataset? the given link is toward your article on the dataset. but dont have any link to download the dataset.

 

Regards

Safi

[Inês Domingues]

Dear Safi,

The link to get INbreast database is:

http://medicalresearch.inescporto.pt/breastresearch/index.php/Get_INbreast_Database.

You will have to email medical...@inescporto.pt asking for it.

[SAFI ULLAH Marwat]

Dear Ines,

 

thank you for your cooperation. Can you please tel lme little bit more about the dataset? wether these imaegs are digitized images or acquired as digital images?

 

Regards,

 

Safi

[Inês Domingues]

Dear Safi,

The images are FFDM, meaning they are NOT digitized. They are
completely digital.

INBREAST has a total of 115 cases (410 images) from which 90 cases are
from women with both breasts affected (four images per case) and 25
cases are from mastectomy patients (two images per case). Several
types of lesions (masses, calcifications, asymmetries, and
distortions) were included. Accurate contours made by specialists are
also provided in XML format.

As your main interest is calcifications, they are present in 301 of
the 410 images. The tag “cluster” was only used in 27 sets of
calcifications, in 21 images (≈1.3 clusters per image). Of these 21
images, only 2 had no single calcifications annotation. A total of
6880 calcifications were thus individually identified in 299 images
(≈23.0 calcifications per image).

Regards,

[SAFI ULLAH Marwat]

Dear Ines,

 

Thank you for the guidance. I need some conversation with you regarding my research. Can you please tell me about the papers which discuss the dtection of individidual microcalcifications only? (I know the fact that individual microcalcifications are of no interest as radiologists look for clusters). I am just looking for publications on individual calcs detection level. your help wil really count on.

 

Warm Regards,

 

Safi

[Inês Domingues]

Dear Safi,

Unfortunately I do not have the time to make you a literature search.
But maybe you can start with the following reviews:

[1] H. D. Cheng, X. Cai, X. Chen, L. Hu, and X. Lou, ‘Computer-aided
detection and classification of microcalcifications in mammograms: a
survey’, Pattern Recognition, vol. 36, no. 12, pp. 2967–2991, 2003.
[2] R. M. Rangayyan, F. Ayres, and J. E. L. Desautels, ‘A review of
computer-aided diagnosis of breast cancer: toward the detection of
subtle signs’, The Journal of Franklin Institute, vol. 344, no. 3–4,
pp. 312–348, 2007.
[3] J. Bozek, M. Mustra, K. Delac, and M. Grgic, ‘A Survey of Image
Processing Algorithms in Digital Mammography’, Recent Advances in
Multimedia Signal Processing and Communications, vol. 231, pp.
631–657, 2009.

and then then some of the references they cite. You can also use the
"Cited by" feature of Google Scholar to check more recent papers.

Regards,


Inês Domingues
-----------------------------------------------------------------------------------------------
PhD student
INESC TEC and Faculdade de Engenharia, Universidade do Porto
https://sites.google.com/site/indezitindezit/
-----------------------------------------------------------------------------------------------