Sperm Analysis

[Taufiq Purna Nugraha]

Dear All,

I'm learning to use image to analyze sperm sample, mainly to measure
head area, head width & length and also tail length.
I know this subject has been discussed quite for some time but I still
have some difficulties.  For the head area I'm using
particle analysis (PA) to measure the area, but the main problem is
how to accurately separate between head and tail (to
measure head area). Because in the PA I define the sperm head as a
circular object with circularity between 0.75-0.9,
I have try several methods to separate the tail from the head using
such as erode and dilate or minimum and maximum filter
and the main problem of this method is the sperm sometimes became to
round since the dilate or maximum filter only
interpolate from the eroded condition.  The best method so far to cut
the tail from the head manually using the pen tools.

Is there anyway to calculate head area without the necessity to
separate head from tail ?  For the head length I'm using ferrets
from ImageJ PA, but for the sperm width (breadth) I still having the
difficulties to count it automatically.  And is there any way to
measure
tail length automatically ? So far I'm using manual methods using line
tools but sometimes became very  frustrating when I have to measure
3000 cells.

For the sample image of the sperm I upload it at
http://picasaweb.google.com/RTP.Nugraha/Sperm/photo#5186514543892861554

Thanks in advance and looking forward to hear any suggestion

--
R. Taufiq Purna Nugraha, DVM.

Laboratory of Reproduction
Div. of Zoology [Museum Zoologicum Bogoriense]
R.C. for Biology
Indonesian Institute of Sciences [LIPI]
Jl. Raya Jakarta – Bogor Km. 46, Cibinong 16911
INDONESIA
Telp : [phone]/64 Ext. 151
Fax : [phone]
e-mail : [hidden email], [hidden email]

[karo03]

Try Morphology plugin "graymorphology, radius 10., circle, close" on  
the gray image. You have than the sperm heads without tail for  
particle analysis.

Regards
Karsten

Am 07.04.2008 um 17:08 schrieb Taufiq Purna Nugraha:
> For the sample image of the sperm I upload it at
> http://picasaweb.google.com/RTP.Nugraha/Sperm/ 
> photo#5186514543892861554

[Taufiq Purna Nugraha]

Dear Karsten and Chris

Thanks for the advice, I try the gray morphology plugins last night, but
this method still have a bias just like the minimum and maximum filter,
where the sperm became to round and over estimate on the head and tail
junction, I will try again on different samples.  And for Chris  I will try
your advice tonight.

btw, is the gray morphology that slow ?, I use it on an image with 2816 x
2112 pixel resolution and it takes more than one minutes to produce the
result.  And do you have any advice how to count the sperm tail length
automatically ?

Cheers,



On Mon, Apr 7, 2008 at 10:44 PM, Karsten Rodenacker <[hidden email]>
wrote:

> Try Morphology plugin "graymorphology, radius 10., circle, close" on the
> gray image. You have than the sperm heads without tail for particle
> analysis.
>
> Regards
> Karsten
>
> Am 07.04.2008 um 17:08 schrieb Taufiq Purna Nugraha:
>
>  For the sample image of the sperm I upload it at
> > http://picasaweb.google.com/RTP.Nugraha/Sperm/photo#5186514543892861554
> >
>

--
R. Taufiq Purna Nugraha, DVM.

Laboratory of Reproduction
Div. of Zoology [Museum Zoologicum Bogoriense]
R.C. for Biology
Indonesian Institute of Sciences [LIPI]
Jl. Raya Jakarta – Bogor Km. 46, Cibinong 16911
INDONESIA
Telp : [phone]/64 Ext. 151
Fax : [phone]
e-mail : [hidden email], [hidden email]

[karo03]

Am 08.04.2008 um 15:43 schrieb Taufiq Purna Nugraha:

> Dear Karsten and Chris
>
> Thanks for the advice, I try the gray morphology plugins last night,  
> but
> this method still have a bias just like the minimum and maximum  
> filter,
> where the sperm became to round and over estimate on the head and tail
> junction, I will try again on different samples.  And for Chris  I  
> will try
> your advice tonight.
>
> btw, is the gray morphology that slow ?, I use it on an image with  
> 2816 x
> 2112 pixel resolution and it takes more than one minutes to produce  
> the
> result.  And do you have any advice how to count the sperm tail length
> automatically ?

Computing time is dependent on the machine and the radius chosen. In  
fact it is slow.
The difference between the original and the closed image allows to  
detect the sperm tails quite well. Possibly with a Threshold, another  
(binary) opening (small) and a skeleton. For each connected object  
(isolated tail) the number of pixels is an estimate, not good but  
possibly sufficient. Another one possibly better is the perimeter/2 of  
each connected skeletonized object.
Karsten

>
> Cheers,
>
>
>
> On Mon, Apr 7, 2008 at 10:44 PM, Karsten Rodenacker <[hidden email]
> >
> wrote:
>
>> Try Morphology plugin "graymorphology, radius 10., circle, close"  
>> on the
>> gray image. You have than the sperm heads without tail for particle
>> analysis.
>>
>> Regards
>> Karsten
>>
>> Am 07.04.2008 um 17:08 schrieb Taufiq Purna Nugraha:
>>
>> For the sample image of the sperm I upload it at
>>> http://picasaweb.google.com/RTP.Nugraha/Sperm/photo#5186514543892861554
>>>
>>
>
>
> --
> R. Taufiq Purna Nugraha, DVM.
>
> Laboratory of Reproduction
> Div. of Zoology [Museum Zoologicum Bogoriense]
> R.C. for Biology
> Indonesian Institute of Sciences [LIPI]
> Jl. Raya Jakarta – Bogor Km. 46, Cibinong 16911
> INDONESIA
> Telp : [phone]/64 Ext. 151
> Fax : [phone]
> e-mail : [hidden email], [hidden email]