High Kernel Times - Batch Processing AVI files

[Cameron Nowell-2]

Hi List ,

I have a whole heap of AVI files I need to batch process from some organ bath experiments and I am hitting a bit of a wall.

A quick bit of background. The user has captured 15 minute videos with a webcam of isolated gut contracting in an organ bath. We want to extract just the area of the gut in the bath and produce a kymogram to show the propagation of contractions along it.

I have the code all written up and it worked fine on a reduced size data set but now that I am processing the full sized movies it all kind of breaks. The movies once opened contain 27,000 frames

The macro loads up an AVI file, applies a pre-defined ROI, crops it out and resaves it elsewhere. The AVI files are opend as virtual stacks to save memory (but even if I open them full it doesn't seem to make a difference).  Once you start processing it duplicates the first video fine but then everything comes to a grinding halt with it processing everything is short bursts of 1000 frames or so. The first video will be done in 10 minutes, the next takes longer and by the third it is taking an hour to do the same thing.

If you go into the task manager and resource manager you can see that the NT Kernel and System if flogging the hell out of everything, which I assume is what is leading to the slow downs.

Is there anything that can be done about this or is this some inherent java issue?

Cheers
Cam



Cameron J. Nowell
Research Facilities Manager

Imaging, FACS and Analysis Core
Monash Institute of Pharmaceutical Sciences
Monash University
399 Royal Parade
(Mail address: 381 Royal Parade)
Parkville, VIC, 3052
Australia

Email: [hidden email]
Mobile: [phone]
Office: [phone]

LinkedIn: Profile
Research Gate:  Profile

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

[Cameron Nowell-2]

Hi List,

Found a solution to my problem. Running it on Linux works perfectly, only takes about 3 minutes to process each video. So it seems to be a windows specific problem.

Cheers

Cam
________________________________________
From: ImageJ Interest Group [[hidden email]] on behalf of Cameron Nowell [[hidden email]]
Sent: Thursday, September 04, 2014 9:30 AM
To: [hidden email]
Subject: High Kernel Times - Batch Processing AVI files

Hi List ,

I have a whole heap of AVI files I need to batch process from some organ bath experiments and I am hitting a bit of a wall.

A quick bit of background. The user has captured 15 minute videos with a webcam of isolated gut contracting in an organ bath. We want to extract just the area of the gut in the bath and produce a kymogram to show the propagation of contractions along it.

I have the code all written up and it worked fine on a reduced size data set but now that I am processing the full sized movies it all kind of breaks. The movies once opened contain 27,000 frames

The macro loads up an AVI file, applies a pre-defined ROI, crops it out and resaves it elsewhere. The AVI files are opend as virtual stacks to save memory (but even if I open them full it doesn't seem to make a difference).  Once you start processing it duplicates the first video fine but then everything comes to a grinding halt with it processing everything is short bursts of 1000 frames or so. The first video will be done in 10 minutes, the next takes longer and by the third it is taking an hour to do the same thing.

If you go into the task manager and resource manager you can see that the NT Kernel and System if flogging the hell out of everything, which I assume is what is leading to the slow downs.

Is there anything that can be done about this or is this some inherent java issue?

Cheers
Cam



Cameron J. Nowell
Research Facilities Manager

Imaging, FACS and Analysis Core
Monash Institute of Pharmaceutical Sciences
Monash University
399 Royal Parade
(Mail address: 381 Royal Parade)
Parkville, VIC, 3052
Australia

Email: [hidden email]
Mobile: [phone]
Office: [phone]

LinkedIn: Profile
Research Gate:  Profile

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

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