Large image

classic Classic list List threaded Threaded
6 messages Options
Reply | Threaded
Open this post in threaded view
|

Large image

Sho Fujisawa
What is the largest size image FIJI can open? By that I mean, the biggest number of pixels in x and y (and I guess z, though not relevant at the moment). Is there any way to circumvent this limit?


Sho Fujisawa
Sr. Digital Microscopist
Molecular Cytology Core Facility
Memorial Sloan-Kettering Cancer Center
417 E. 68th St.
Zuckerman Room 1834
New York NY 10065
(212) 639-6667
(646) 888-2186


=====================================================================

     Please note that this e-mail and any files transmitted from
     Memorial Sloan Kettering Cancer Center may be privileged, confidential,
     and protected from disclosure under applicable law. If the reader of
     this message is not the intended recipient, or an employee or agent
     responsible for delivering this message to the intended recipient,
     you are hereby notified that any reading, dissemination, distribution,
     copying, or other use of this communication or any of its attachments
     is strictly prohibited.  If you have received this communication in
     error, please notify the sender immediately by replying to this message
     and deleting this message, any attachments, and all copies and backups
     from your computer.

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html
Reply | Threaded
Open this post in threaded view
|

Re: Large image

Rasband, Wayne (NIH/NIMH) [E]
On Jan 14, 2016, at 9:01 AM, Sho Fujisawa <[hidden email]> wrote:
>
> What is the largest size image FIJI can open? By that I mean, the biggest number of pixels in x and y (and I guess z, though not relevant at the moment). Is there any way to circumvent this limit?

2D images are contained in Java arrays, which have a maximum size of 2^31 (2 gigapixels). A square 2 gigapixel image is 46341 x 46341. I do not know of a way to circumvent this limit.

-wayne

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html
Reply | Threaded
Open this post in threaded view
|

Re: Large image

Mike Ingle
If the maximum native image size is insufficient, then you should be able
to write an image opener which opens the image as an (n x m) array of
tiles, where each tile is within the native image size limits.

On Fri, Jan 15, 2016 at 6:22 AM, Rasband, Wayne (NIH/NIMH) [E] <
[hidden email]> wrote:

> On Jan 14, 2016, at 9:01 AM, Sho Fujisawa <[hidden email]> wrote:
> >
> > What is the largest size image FIJI can open? By that I mean, the
> biggest number of pixels in x and y (and I guess z, though not relevant at
> the moment). Is there any way to circumvent this limit?
>
> 2D images are contained in Java arrays, which have a maximum size of 2^31
> (2 gigapixels). A square 2 gigapixel image is 46341 x 46341. I do not know
> of a way to circumvent this limit.
>
> -wayne
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html
Reply | Threaded
Open this post in threaded view
|

Re: Large image

Piotr Wendykier-2
In reply to this post by Rasband, Wayne (NIH/NIMH) [E]
On Fri, Jan 15, 2016 at 6:22 AM, Rasband, Wayne (NIH/NIMH) [E]
<[hidden email]> wrote:
> On Jan 14, 2016, at 9:01 AM, Sho Fujisawa <[hidden email]> wrote:
>>
>> What is the largest size image FIJI can open? By that I mean, the biggest number of pixels in x and y (and I guess z, though not relevant at the moment). Is there any way to circumvent this limit?
>
> 2D images are contained in Java arrays, which have a maximum size of 2^31 (2 gigapixels). A square 2 gigapixel image is 46341 x 46341. I do not know of a way to circumvent this limit.
>

That is why I have implemented JLargeArrays library:
https://gitlab.com/ICM-VisLab/JLargeArrays

Piotr

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html
Reply | Threaded
Open this post in threaded view
|

Re: Large image

ctrueden
In reply to this post by Sho Fujisawa
Hi Sho,

> What is the largest size image FIJI can open? By that I mean, the
> biggest number of pixels in x and y (and I guess z, though not
> relevant at the moment).

As Wayne says, there is a 2 gigapixel limit when opening and displaying
image planes using the ImageJ (1.x) user interface's default image viewer.

> Is there any way to circumvent this limit?

One major goal of the ImageJ2 project is to break this limit.

ImageJ2 uses the ImgLib2 library [1] as its data model, which has much
larger limits on the number of pixels (theoretically: ~2^63 per dimension,
up to ~2^31 dimensions, totaling ~2^63^31 pixels).

Hence, you can open larger images using the ImageJ2 API. But you will not
be able to display them in the user interface yet.

The SCIFIO library [2] also supports opening such images on-demand, so that
blocks are read from disk as you iterate over the image. This is similar to
(but more powerful than) ImageJ 1.x's virtual stacks feature.

In this way, you can write code to process these large images without
displaying them.

The Fiji project [3] already includes the BigDataViewer (BDV) plugin [4],
which currently functions as an alternative viewer, to display arbitrarily
large images backed by ImgLib2. In the future we hope to integrate
BDV-driven UI technology more completely into ImageJ core. But there are a
couple of limitations right now:

1) BDV was originally designed for large SPIM data; opening large image
files more generally currently requires some coding. But we want to change
this.

2) Many ImageJ 1.x plugins assume the data is stored in an ImageJ 1.x data
structure, which is not what BDV uses. So "mixing and matching" IJ1 and IJ2
functionality is tricky here.

But we are working to lift these restrictions as time goes on.

Regards,
Curtis Rueden, ImageJ2 architect
Tobias Pietzsch, BigDataViewer architect

[1] http://imagej.net/ImgLib2
[2] http://imagej.net/SCIFIO
[3] http://imagej.net/Fiji
[4] http://imagej.net/BigDataViewer

On Thu, Jan 14, 2016 at 3:01 PM, Sho Fujisawa <[hidden email]> wrote:

> What is the largest size image FIJI can open? By that I mean, the biggest
> number of pixels in x and y (and I guess z, though not relevant at the
> moment). Is there any way to circumvent this limit?
>
>
> Sho Fujisawa
> Sr. Digital Microscopist
> Molecular Cytology Core Facility
> Memorial Sloan-Kettering Cancer Center
> 417 E. 68th St.
> Zuckerman Room 1834
> New York NY 10065
> (212) 639-6667
> (646) 888-2186
>
>
> =====================================================================
>
>      Please note that this e-mail and any files transmitted from
>      Memorial Sloan Kettering Cancer Center may be privileged,
> confidential,
>      and protected from disclosure under applicable law. If the reader of
>      this message is not the intended recipient, or an employee or agent
>      responsible for delivering this message to the intended recipient,
>      you are hereby notified that any reading, dissemination, distribution,
>      copying, or other use of this communication or any of its attachments
>      is strictly prohibited.  If you have received this communication in
>      error, please notify the sender immediately by replying to this
> message
>      and deleting this message, any attachments, and all copies and backups
>      from your computer.
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html
Reply | Threaded
Open this post in threaded view
|

Re: Large image

Albert Cardona-2
In reply to this post by Sho Fujisawa
Imglib2 offers the CellImg that abstracts over a mosaic of N-dimensional tiles.

TrakEM2 will let you display and montage 2^31 2-dimensional tiles.

Both are available in Fiji.

Albert

> On Jan 14, 2016, at 9:01 AM, Sho Fujisawa <[hidden email]> wrote:
>
> What is the largest size image FIJI can open? By that I mean, the biggest number of pixels in x and y (and I guess z, though not relevant at the moment). Is there any way to circumvent this limit?
>
>
> Sho Fujisawa
> Sr. Digital Microscopist
> Molecular Cytology Core Facility
> Memorial Sloan-Kettering Cancer Center
> 417 E. 68th St.
> Zuckerman Room 1834
> New York NY 10065
> (212) 639-6667
> (646) 888-2186
>
>
> =====================================================================
>
>     Please note that this e-mail and any files transmitted from
>     Memorial Sloan Kettering Cancer Center may be privileged, confidential,
>     and protected from disclosure under applicable law. If the reader of
>     this message is not the intended recipient, or an employee or agent
>     responsible for delivering this message to the intended recipient,
>     you are hereby notified that any reading, dissemination, distribution,
>     copying, or other use of this communication or any of its attachments
>     is strictly prohibited.  If you have received this communication in
>     error, please notify the sender immediately by replying to this message
>     and deleting this message, any attachments, and all copies and backups
>     from your computer.
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html

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