About re-processing images acquired with old versions

Question from Alison Cawood :
Hi Marc. I'm having issues with doublettes again. I have reprocessed
images using the 5.02 version of Zooprocess and I have tried processing
converting from the 16 bit and the converted 8 bit images. Not sure what
to do about this.
Thanks,
Alison

Dear Alison,
I is very hard to ansxer such question containing no detail... I try to add some information :
Two options :
1) You have set Zooprocess to save a REAL RAW image in the "Zooscan_scan\_raw" folder when you scanned your samples. You can thus use the CONVERSION tool and re-process everything (probably using rolling ball because you don't have any background scanned). Carefull, the images in the "Zooscan_scan" will be replaced. Save them in a separate folder if they haven't been zipped in the "_zip" folder during your process.
2) You don't have saved these RAW images : the new Zooprocess 5.02 will not make any improvement in your image quality. You don't need to reprocess.
Regards
Marc

This problem occurred even when converting and reprocessing the raw images.

What details would be helpful to you in addressing this issue? These images are tagged differently (some as 1 and some as 2), but they are both in the pid files.

Dear Alison,
What do you call "raw" images ? Again my question, have they been saved in the "_raw" folder by Zooprocess or in the "zooscan_scan" folder ?
In the first case, you can use the standard conversion tool. In the later you don't need to re-process anything !
Can you make a copy of your monitor screen with Windows explorer opened and post it ?
Regards
Marc P.

Details : opertaration performed ? For what purpose ? Problem encountered ? Copy of error messages....

I am using images from the _raw folder. I am attempting to batch process. From the initial zooprocess screen, I select my project and "convert and process images and organisms in batch mode." In the Process (Converted) Images box, I choose "convert and process image and particles (images in raw folder)" as the select process option. The image process works well. However, in the PID file for these images, there are doublettes. I noticed them when I extracted the images. I am finding about 150 pairs of doublettes per 5000 or so thumbnails. This is important because it changes the abundances of organisms. I do not believe that this is a problem that I can fix. I think that it is something that has to be changed within zooprocess.

Dear Alison,

The version 3 of Zooprocess automatically removed the duplicates (tagged) objects from the PID file of the frame "_2" while they did remain in the MEAS.txt files. The later versions do not remove them because we now think that it is safer to keep all the data and it would be easy in MATLAB, Excel, R... to sort on that criteria and remove them. Zooprocess DO NOT extract vignettes of objects tagged "2" (except bug ???).

It is thus normal to find duplicates in the PID files and they are tagged "2" in the last column of the files.
These duplicates should not be extracted in vignettes in any of the tools of Zooprocess.

Are your 150 pairs of organisms tagged "2" or not ? This is the important question to detect a "bug".
Did Zooprocess extracted two vignettes for them (one from each frame ?).
Can you check this in detail and tell me also which tool was used to extract the vignettes ?

Hope this is a begining of answer... I now understand better your problem. Thank you !

Regards?

Marc P.

There is an object from each half image (_1 and _2) and they are tagged differently. I am extracting the images from a text file based on the prediction. The text file is obtained via Plankton Identifier/Tanagra. I also see duplicate images when I use the extract all vignettes option of the extract vignettes for plankton identifier and when selecting images for manual measurements (although I haven't done this often, so I don't know how persistent this problem is with these tools).

Dear Alison,
I do now understand... It is a bug that is fixed in the new version to be released. I will post it on the web next monday for all users.
A apologize for this problem.
You said that you have "extracted vignettes from prediction" using the text files issued from PKId. I think that you will only have to redo this operation with the new version. You will get only vignettes of objects tagged "1" and they will be placed in a new subfolder of the "sorted vignettes" folder.
Please wait next week !
Regards
Marc P.