standard deviation in an analysed image

[sebastian]

I am really curious to understand the utility of the standard deviation shown in fiji. i am a bit confused, as , if i have to show average integrated density of a single cell in a population, do i need to take average of the standard deviation shown in fiji or do i need to calculate std deviation of my own. because if i calculate my own std dev, then because of high heterogeneity of the system my std dev comes more than the estimated integrated density , which thereby disapproves with the t test of the sample. please help me in understanding how do i represent my data.

[Herbie-4]

Good day Sebastian,

could you please be more specific.

What do you mean by:
"take average of the standard deviation"

What are you trying to do?
Please give a step by step description and perhaps show a typical image.

Looking forward

Herbie

:::::::::::::::::::::::::::::::::::
On 06.02.15 06:12, sebastian wrote:

> I am really curious to understand the utility of the standard
> deviation shown in fiji. i am a bit confused, as , if i have to show
> average integrated density of a single cell in a population, do i
> need to take average of the standard deviation shown in fiji or do i
> need to calculate std deviation of my own. because if i calculate my
> own std dev, then because of high heterogeneity of the system my std
> dev comes more than the estimated integrated density , which thereby
> disapproves with the t test of the sample. please help me in
> understanding how do i represent my data.
>
>
>
> -- View this message in context:
> http://imagej.1557.x6.nabble.com/standard-deviation-in-an-analysed-image-tp5011470.html
>
>

Sent from the ImageJ mailing list archive at Nabble.com.
>
> -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[sebastian]

dear herbie, good day.

my question is, when i take a stacked image through confocal for the cells. what i do is, i try to see the max intensity of the stack and then create an roi around each cells ( for eq i have 20 cells in an image) and for each cells i perform measure. this result gives me a standard deviation and integrated density along with many other measures for 20 cells. this standard deviation what i understand is the deviation of all the pixels from the mean gray value for a cell. therefore in total if i consider all the 20 cells and i bypothetically calculate an average integrated density for the expression of my protein in one cell as it is a highly hetergenous system. hence my question is, do i need to take an average of the standard deviation shown in this steps for 20 cells and mark that in my plot for the integrated density of the image or do i need to recalculate the deviation manually in 20 cells. the problem is if i manually calculate the deviation, the result is tooo high for the standard deviation as it is a highly heterogenous system. but the p value in between the result is significant so if i show high deviation, is it correct? what standard deviation has to be plotted in the graph for integrated density of the image.

Regards,
Sebastian John

[Herbie-4]

Dear Sebastian,

if "it is a highly hetergenous system" then you are to show 20 values
with their respective Standard Deviations. In such cases, averaging over
the ROIs is definitely no good practice.

Best

Herbie

:::::::::::::::::::::::::::::::::::
On 07.02.15 06:09, sebastian wrote:

> dear herbie, good day.
>
> my question is, when i take a stacked image through confocal for the cells.
> what i do is, i try to see the max intensity of the stack and then create an
> roi around each cells ( for eq i have 20 cells in an image) and for each
> cells i perform measure. this result gives me a standard deviation and
> integrated density along with many other measures for 20 cells. this
> standard deviation what i understand is the deviation of all the pixels from
> the mean gray value for a cell. therefore in total if i consider all the 20
> cells and i bypothetically calculate an average integrated density for the
> expression of my protein in one cell as it is a highly hetergenous system.
> hence my question is, do i need to take an average of the standard deviation
> shown in this steps for 20 cells and mark that in my plot for the integrated
> density of the image or do i need to recalculate the deviation manually in
> 20 cells. the problem is if i manually calculate the deviation, the result
> is tooo high for the standard deviation as it is a highly heterogenous
> system. but the p value in between the result is significant so if i show
> high deviation, is it correct? what standard deviation has to be plotted in
> the graph for integrated density of the image.
>
> Regards,
> Sebastian John
>
>
>
> --
> View this message in context: http://imagej.1557.x6.nabble.com/standard-deviation-in-an-analysed-image-tp5011470p5011496.html
> Sent from the ImageJ mailing list archive at Nabble.com.
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html