Greetings.
As I mentioned in my previous posts, I'm starting in the use of ImageJ and I recently install the plugin Fast Morphology ( http://imagejdocu.tudor.lu/doku.php?id=plugin:morphology:fast_morphological_filters:start). As also mentioned in my latest post, even though the plugin works well (morphological filters are apparently more rapid) in the status bar does not show any information about the time of execution of the filters, so not to exactly how long they take. In a response to my previous post someone told me that in fact the Morphology Fast plugin shows the execution time of the operators in the status bar (as I said in my case nothing is displayed). For this reason I decided to ask regarding this plugin (Fast Morphology: http://imagejdocu.tudor.lu/doku.php?id=plugin:morphology:fast_morphological_filters:start). Does it show or not the execution time of the operators in the status bar?. I would like someone who has installed and tested the plugin on your computer tell me if you have the same problem as me or if instead the plugin works as expected (ie, as said the person who answered my previous post, the plugin execution times shown in the status bar). If the answer is the second option, does that mean that I install the plugin 'Fast Morphology' or install ImageJ incorrectly or not? ... What could do in that case?. The person who answered my previous post (whom I thank), also advised me to write a macro to measure the execution time of the operators myself. So, I've been learning about the ImageJ macro language and wrote an macro to solve my problem. The macro file I wrote (which attached with this post), it's just for a test. However, I wanted to know if the way I'm measuring the execution time is correct and if there is a better way (more precise). As always, thanks in advance for any help and / or suggestions. -- Isaac Perez -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
On Sunday 30 Jun 2013 10:35:08 you wrote:
> The person who answered my previous post (whom I thank), also advised > me to write a macro to measure the execution time of the operators > myself. So, I've been learning about the ImageJ macro language and > wrote an macro to solve my problem. The macro file I wrote (which > attached with this post), it's just for a test. However, I wanted to > know if the way I'm measuring the execution time is correct and if > there is a better way (more precise). If you do not see anything in the status bar, but you can see it with other commands, then the plugin does not output that info. Not all commands do. I am not sure what are you trying to do with this, but have you considered that the algorithms you are comparing might rely on image contents? For example there could be some condition testing, like a shortcut where if there are no "dilatable" pixels the plugin does not compute anything or something along those lines. If so then the test that you are trying to do might not be particularly useful as it will depend on the image you are using for benchmarking. If you could explain what is the need for the comparisons then others might be able to help a bit more. Also remember that speed of execution relies on a number of things as well. See the benchmak test in the ImageJ wiki. Regards Gabriel -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
Greetings.
Thanks for your reply Gabriel. What I wanted know is if is normal or not that the plugin (Fast Morphology: http://imagejdocu.tudor.lu/doku.php?id=plugin:morphology:fast_morphological_filters:start) not show that information on the bar of status. On the other hand, for the measurement of the time of execution, the macro I wrote that basically does the following: // Open an image t1 = getTime(); run("Morphological Filters", "operation=[White Top Hat] element=Square radius=31"); t2 = getTime(); // T1 and t2 used to calculate the execution time. // Write the results to a file Well, I also wrote the macro like above but for the command 'Process> Filters> Minimum...', and (since this command shows the execution time in the status bar), I realized that the estimated runtime in the macro is not exactly the same as the execution time shown in the status bar (the difference varies between 10 and 50 msec). For that wondering: is there a better way (more precise) to measure the execution time of a command? Thanks in advance for any help and / or suggestions -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
In reply to this post by Isaac Perez
Hi Isaac,
I'm the author of the plugin 'fast morphology' ; I recently joined the list, and noticed you have some questions about it, so I suppose I should reply... First of all, there was an update of the plugin in the beginning of June. One of the modifications consisted in using the PluginFilter interface. This results in the display of the execution time in the status bar. If you downloaded an earlier version, no execution information should be displayed. Therefore I would recommend updating the plugin. About the plugin itself: it uses two kinds of acceleration. One is to decompose structuring elements into linear structuring elements. Basic operations such as dilation or erosion can be applied directly on the input image (without creating a new image for result). Second improvement is to implement linear erosions and dilations using moving histogram technique. The principle is to update the local min or local max only when necessary. Current implementation uses a circular buffer. Running time is acceptable, but it is still possible to improve. For binary images, faster implementation can be envisioned. For local histograms, faster algorithms exist (ie, von van Herk or Gil and Werman), but I haven't taken time to implement them. Anyway, feel free to ask precisions if you need. Regards, David -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
Thanks for your answer David Legland.
Where do I download the new plugin or rather the plugin upgrade? ... In the exact same direction as the previous version (http://imagejdocu.tudor.lu/doku.php?id=plugin:morphology:fast_morphological_filters:start) or another address? How to install? -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
In reply to this post by Isaac Perez
Hi Isaac,
the new plugin is located at the same location. You have given the link to get the new plugin. There is a link to download the last version (currently 1.3), and instructions on how to install it (like most other plugins: simply copy the jar file into the plugins directory). Regards, David -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
Free forum by Nabble | Edit this page |