X-Y coordinates to pixels

[Brian Sellar]

Leslie,

I'm jumping in on this thread so forgive me if I'm off the mark.

If you record an image of something then the amount of information you
can gain depends on how many pixels cover the area of the thing you
photographed.

If you were taking an image of a complete A4 piece of paper for example
which is about 300mm x 210mm and your camera produced images of 720 x
560 pixels then to find the image lengthscale compared to realworld you
divide the length of object (paper)by 720 and width of object by 560.
So every pixel in your image represents ~0.4mm top to bottom and across.

So a pixel at pixel location (21,47) in your XY coords would be
[21x0.4mm] along and [47x0.4mm] down from one corner of the image
(depending on where the program starts its rows and columns and where
on the pixel the XY coord represents (top left corner??) This
represents a real world length. Now you don't have integer lengths (the
displacement from the top corner of the paper to your point of interest
would be 20.591mm). But since pixels are the smallest element in an
image and are in rows and columns they'll always be whole numbers.

Brian

[Jerome Mutterer]

Getting the mouse (x,y) location will always return you pixel  
coordinates or coordinates in units specified in set scale dialog. If  
you want to achieve more precision, you can imagine resampling your  
image with the (image adjust size) dialog that will not affect the  
scale parameter. Then you'll get smaller pixels (in scaled unit) and  
more precision with your measurements. However, resampling can  
introduce artifacts in the form of pixels with values that do not  
exist in your original image, or even significantly stretch or shrink  
structures when their size is close to the size of single pixels. So  
you might also loose some precision there.

Jerome

Le 3 févr. 06 à 15:12, Leslie Day a écrit :

> Okay. Then I rephrase my question... How can I get image J to give me
> another measurement of a point, besides pixels. Some that is much more
> precise and can help me measure small amounts of movement?
>
> I am not sure what a Gaussian is.
>
> Leslie Day
> Biology Department
> Northeastern University
> [hidden email]
> 617-373-3780
>
> -----Original Message-----
> From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of
> Michael Cammer
> Sent: Thursday, February 02, 2006 9:37 AM
> To: [hidden email]
> Subject: Re: X-Y coordinates to pixels
>
> To measure particles, either you need to do it per pixel which, by
> definition, is a discrete number, or you need to look at a block of  
> pixels
> and do a gaussian or other fit if you need to find a fraction of a  
> pixel
> location.
> -mc
>
>
>
> At 09:14 AM 02/02/06 -0500, you wrote:
>> Dear Sir,
>>
>> I have the same problem. Could I calculate the displacement of two
>> particles according to the pixel size?
>>
>>
>>
>>
>>
>>
>>
>> **********************************************
>> Dongfang Liu
>> Laboratory of Immunogenetics
>> National Institute of Allergy and Infectious Disease
>> National Institute of Health
>> Twinbrook II, Room 207
>> 12441 Parklawn Drive
>> Rockville, MD 20852
>> Phone: [phone]
>> Fax:[phone]
>> email:[hidden email]
>> **********************************************
>>
>>> ----------
>>> From:         Curtis Rueden
>>> Reply To:     List IMAGEJ
>>> Sent:         Wednesday, February 1, 2006 3:20 PM
>>> To:   List IMAGEJ
>>> Subject:      Re: X-Y coordinates to pixels
>>>
>>> Hi Leslie,
>>>
>>> I do not know if there is a way to measure points at higher  
>>> precision in
>>> ImageJ (though I would be surprised if there is not), but this is  
>>> one of
>>> the things VisBio (http://www.loci.wisc.edu/visbio/) can do. You can
>>> place markers in double-precision floating point, and then export  
>>> them
>>> to a text file for use in a spreadsheet. See the Overlays page in  
>>> the
>>> "Data Transforms" section of the program's built-in help for  
>>> details.
>>>
>>> -Curtis
>>>
>>> Leslie Day wrote:
>>>
>>>> I am hoping someone could help me. I am trying to map a point as it
> moves
>>>> through a series of images in order to find out velocity of the  
>>>> object
> and
>>>> distance traveled.. I figured the best way would be to keep  
>>>> creating
>> points
>>>> with the point selector and then map the points. However, currently
> when I
>>>> use the point selector it gives me x-y coordinates in whole  
>>>> numbers. I
>> need
>>>> something with more significant figures (3 decimal spots) to be  
>>>> able
>> to see
>>>> the movement. Maybe in pixels or ?m. Perhaps there is another  
>>>> way to
>> detect
>>>> linear velocity and distance traveled through a series of images.
>>>>
>>>>
>>>>
>>>> Thanks
>>>>
>>>>
>>>>
>>>> Leslie Day
>>>>
>>>> Biology Department
>>>>
>>>> Northeastern University
>>>>
>>>> [hidden email]
>>>>
>>>> 617-373-3780
>>>>
>>>>
>>>
>
> ______________________________________________________________________
> ______
> Michael Cammer   Analytical Imaging Facility   Albert Einstein  
> Coll. of Med.
> Jack & Pearl Resnick Campus      1300 Morris Park Ave.     Bronx,  
> NY  10461
> [phone]       Fax:  [phone]      URL:  http://
> www.aecom.yu.edu/aif/
>    **This electronic transmission contains information that is  
> privileged.**