StarNet - new module for star removal

[acmalko]

It works well, thanks

Tried 128 et 8 size stride. The last one was very time consuming 

I've re-read the instructions written in the stand-alone version to remember what "stride" meant
Perhaps add those instructions in the readme file of the pi version

Good job and thank you

[dld]

I thought that you just stretch the image, remove stars and then 'destretch' result back into its original form. Then, assuming that all the nebulosity is dim, you would get good image.

Hello Nikita, this is a great point!

For a starless region and an invertible MTF, the InverseMTF of MTF(starless) should return starless Thus we can subtract the InverseMTF(starless) from the linear image and obtain the stars. No need to complicate things and trying to solve a more difficult mathematical problem!

I'll have to find some time to work on this. In the meantime, any input is always welcome!

[praspekt]

just tried to install starnet and it's not coming up when I search for it in pixinsight install modules. when I do an api check, everything is api version 0x162. i'm running a mac.

any advice?

thanks

[pfile]

it's windows-only at this point (unless something changed since i downloaded the zipfile a while back.)

rob

[sfacchin]

Hi Nikita,
I tried to install the module, but PI does not recognize it. I dowlnoaded the Starnet_PI folder and copied its contents into the folder C:\Program Files\PixInsight\bin.
I tried the command Install Module on the PI, the following messages appeared on a snall information window:

0 additional PixInsight module(s) were found on directory:
C:/Program Files/PixInsight/bin.

I'm running on windows 764 bits Ultimate up to date and with the last PI release.
Could you help me?

Thanks in advance for your support and best regards,
Sergio (Venice Italy)

[John Massey]

I had the same issue, my cpu is to old. reference the readme text file that came with Starnet.
"Second, the tensorflow.dll was compiled to support AVX instructions
and will only work if your CPU supports these AVX instructions.
Which means that CPUs made before 2011 will not work."

[praspekt]

it's windows-only at this point (unless something changed since i downloaded the zipfile a while back.)

rob

Thanks! I was so excited about it, I must have skipped over that part. would be amazing if it could be made for Mac!!!

[pfile]

you may be able to download a tensorflow.dll compiled without AVX instructions from tensorflow.org. if not, the only other option would be to recompile tensorflow from scratch for your CPU. it will be slow, but it will probably work.

https://tensorflow.org

rob

[pfile]

it's windows-only at this point (unless something changed since i downloaded the zipfile a while back.)

rob

Thanks! I was so excited about it, I must have skipped over that part. would be amazing if it could be made for Mac!!!

i think nikita is working on that, but had some trouble with it.

rob

[pbkoden]

Thank you for this. Your add-in works amazingly and this is a great native process addition to Pixinsight. I will be keeping an eye on this thread for further development. Some things I would love to see at some point in the future:

• progress reporting to the process console. Especially for large images, the process can take a long time. Knowing it is working and not locked up would be nice
• ability to work on linear images
• ability to train the program on our own data (within pixinsight)

But great work!

[sfacchin]

 Hi,
probably my CPU is to old. In the weekend I'll try with a newer one.

     Thanks and best regards to all,
     Sergio

[dave_galera]

Hi Nikita,

I have downloaded the Mac version and it works perfectly with your supplied test .tif files, however, when I convert a .jpg file to a .tif file in Photoshop it does not work.......it goes through the motions and says done but does not save the processed image.....any ideas?

Thanks,
Dave

This is the console log:
Daves-MBP:StarNet_MacOS davewatson$ /Users/davewatson/Desktop/StarNet_MacOS/run_rgb_starnet.sh /Users/davewatson/Desktop/StarNet_MacOS/test.tif
Reading input image... Done!
Bits per sample: 16
Samples per pixel: 3
Height: 712
Width: 1048
Restoring neural network checkpoint... Done!
2019-06-27 22:17:59.934499: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.2 AVX AVX2 FMA
Total number of tiles: 204
100% finished

Done!

[dave_galera]

OK I sorted it the tif file was 8 bit and not 16 bit

Sry, Dave

[dave_galera]

My bad, I should have also congratulated Nikita on a superb piece of software....thanks

Dave

[sfacchin]

 Hello to all,
I tried again on the I5 processor, and the followoing error appeared.
Any idea? Could someone help me?

   Thanks a lot and best regards,
   Sergio

StarNet: Processing view: NGC7023_1920_1080
Writing swap files...
239.349 MiB/s
Starting star removal procedure...
Stride: 8
Image size: 1080x1920
Number of channels: 3
Color space: RGB
Bits per sample: 16
Has alpha channels: false
Float sample: false
Restoring neural network checkpoint...
 Error: Checkpoint file not found!
Reading swap files...
290.779 MiB/s
<* failed *>