[MINC-development] volume_stats/mincstats???

[Robert VINCENT]

Hi,

I've taken a look at this.  If I follow Peter's comments and the program's
logic, it looks as though the following line needs to be added to
mincstats.c:

Index: mincstats.c
===================================================================
RCS file: /software/source/minc/cvsroot/minc/progs/mincstats/mincstats.c,v
retrieving revision 1.11
diff -c -r1.11 mincstats.c
*** mincstats.c	5 Sep 2002 00:41:57 -0000	1.11
--- mincstats.c	12 Aug 2003 18:09:54 -0000
***************
*** 618,623 ****
--- 618,624 ----
              else
                 stats->median = ((double)median_bin + (0.5 - cdf[median_bin])
                                  * pdf[median_bin + 1]) *  hist_sep;
+ 	    stats->median += hist_centre[0];

              stats->majority = hist_centre[majority_bin];
              stats->biModalT = hist_centre[bimodalt_bin];


Does this look correct?

	-bert

On Tue, 12 Aug 2003, Jacqueline Chen wrote:

>
> Hello!
>
> Peter Neelin suggested that I email my problem here...hope you can help!
>
> My data is here:
>
> /home/bic/chen/mincstats_dir
>
> the commands I used were:
>
> volume_stats lipst10_MT_to_e2.mnc.gz
> and
> mincstats lipst10_MT_to_e2.mnc.gz
>
> I have found a bug...specifically with the calculation of the median.
>
> Peter's message:
>
> "Send your message, along with the location of the file and the
> command-line to minc-development@bic.mni.mcgill.ca. Either Steve or Andrew
> should fix this. You can tell them that I said that the median calculation
> looks like it is not adding the histogram offset (centre of the zeroth
> bin). Someone should check all of the results when using a volume that has
> a significantly non-zero minimum (I suspect that Andrew, who wrote the
> code, did not use such a volume for testing)."
>
> here are my outputs:
>
> =====================
> volume_stats output
>
> File:         lipst10_MT_to_e2.mnc.gz
> # voxels:     2949120
> % of total:   100
> Volume (mm3): 8.84736e+06
> Min:          -114.595
> Max:          85.3346
> Sum:          1.57703e+07
> Sum2:         4.93394e+08
> Mean:         5.34745
> Variance:     138.707
> Stddev:       11.7774
> ***************Median:       -0.00774425******************
> Majority:     -0.00774425
> BiModalT:     15.2738
> PctT (-1.79769e+308):    -114.619 85.359
> Entropy:      3.13281
> CoM_voxel:    zspace:27.4898 yspace:135.843 xspace:96.4351
> CoM_world:    zspace:-29.6355 yspace:12.7787 xspace:1.77514
>
> =============
> mincstats output
>
> File:              lipst10_MT_to_e2.mnc.gz
> Mask file:         (null)
> Total voxels:      2949120
> # voxels:          2949120
> % of total:        100
> Volume (mm3):      8847360
> Min:               -114.5949692
> Max:               85.33459454
> Sum:               15783756.65
> Sum^2:             493391934.2
> Mean:              5.35202252
> Variance:          138.6573119
> Stddev:            11.77528394
> CoM_voxel(z,y,x):  27.49728313 135.8338074 96.43413711
> CoM_real(x,y,z):   1.777973285 12.78375223 -29.61186191
>
> Histogram:         (null)
> Total voxels:      2949120
> # voxels:          2949119
> % of total:        99.99996609
> ****************Median:            114.4969327***********
> Majority:          0.01465323195
> BiModalT:          84.88475037
> PctT [  0%]:       0
> Entropy :          2.815999481
>
> ============================
>
> many thanks,
>   Jacqueline
>
> _______________________________________________
> MINC-development mailing list
> MINC-development@bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-development
>