how many input photos are required to make a good mosaic? #8

Open
bluekite2000 opened this Issue Sep 7, 2011 · 1 comment

Comments

Projects
None yet
2 participants

how long does the program take? does the program take a longer time if the input images have large size or? What other factors affect the program's speed or the quality of the final mosaic?

Owner

ideamonk commented Jan 11, 2012

I've observed the following -

  • source images with less color variations make good mosaics
  • hence grayscale, b/w or desaturated images make good mosaics
  • for colorful source images, try have a big repository of thumbnails
  • I haven't benchmarked time performance yet. The time majorly depends on size of source image and zoom factor.
    • searching for colors is optimized via caching
    • if you face long times on very big collection of thumbnails, try installing the rtree python module, it would speed up the color to thumbnail queries
  • for a web app, I'd suggest keeping zoom factor low to produce images 1.5 or 2 times bigger than source
  • also try turning it into a long running task instead of doing it from django directly because it may take a time longer than http timeout
  • I feel celery might be useful here http://www.slideshare.net/idangazit/an-introduction-to-celery http://www.slideshare.net/ericholscher/large-problems
  • also if images turn to have distinguishable patches where same thumbnail gets repeated all over, try increazing the fuzz factor

Thanks for using it in a web app. How has been your experience so far? Would love to know what kind of server are you running it on and if this has been consuming too many resources (cpu+mem) :) ?

@ghost ghost assigned ideamonk Jan 11, 2012

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment