PAT1 CT data statistical information

[tatsuaki]

Hi, 

I'm trying to develop radiomics tool.

Now, I can get the same result with IBSI reference when using digital phantom1.

However, when using CT, I cannot reproduce IBSI reference results.

In my program, PAT1 data's mean value(3D, roi mask used) is -46 (HU) without removing outliers.

I would like to check this value is correct or not.

And, this is another question,

When removing outliers, I used "mean + zscore*SD" for the maximum and "mean - zscore*SD" for the minimum.

Here, mean is a mean value of voxels(only used voxels located at roi mask), zscore is a value to determine CI range(e.g, zscore 2.58 means 99%CI), SD is a standard deviation of voxels ( only used voxels located at roi mask ).

In the IBSI reference manual Table 5.1 Config. D, shows an example of outliers 3 sigma.

In this case, Intensity Based Statistical Feature's Maximum value is  521 HU and Intensity Based Statistical Feature's Minimum value is  -724 HU.

But I can not reproduce these values (My results max:723 min:-964 using above remove outliers calculation). 

I appreciate it that if you let me know your results of mean, max, min values (no remove outlier and no normalize) about PAT1.

(Sorry, I am not good at English..)

Tatsuaki

[tatsuaki]

Hi, 

After that, I realize that my setting was wrong. I forgot resampling to [2,2,2] before calculation.

Now, I get the correct mean value (-49) under the configuration of C (in Table 5-1).

Thanks.

2022年3月4日金曜日 14:58:56 UTC+9 tatsuaki: