User:Nrp/Weekly updates/20080712: Difference between revisions

From OLPC
Jump to navigation Jump to search
(New page: == Saturday == * Performance testing on largest connected component. It runs at a usable speed on the XO, and about twice as fast as the other connected components function in pygame. == ...)
 
No edit summary
 
(3 intermediate revisions by the same user not shown)
Line 3: Line 3:
== Sunday ==
== Sunday ==
* Started writing a function to get the average color of a surface or a part of a surface.
* Started writing a function to get the average color of a surface or a part of a surface.
== Monday ==
* Finished average color function, wrote an example script to show its use.
== Tuesday ==
* Worked on improving the performance of functions that access surfaces.
== Wednesday ==
* Finished performance improvements: average_color is 4x faster, threshold is 2x faster, and mask.from_surface is 2x faster when using per pixel alphas.
* Documented connected_component a little better
== Thursday ==
* Added support for a second surface back into threshold
* Applied the same performance improvements to average_surfaces
* Got rid of the gcc warnings that I could get rid of
== Friday ==
* First attempts at integrating my camera stuff with [[Physics]]
* Wrote an example script that shows the basic process of combining get_image, colorspace, average_color, threshold, connected_component, and centroid to do hue based object tracking.

Latest revision as of 05:41, 12 July 2008

Saturday

  • Performance testing on largest connected component. It runs at a usable speed on the XO, and about twice as fast as the other connected components function in pygame.

Sunday

  • Started writing a function to get the average color of a surface or a part of a surface.

Monday

  • Finished average color function, wrote an example script to show its use.

Tuesday

  • Worked on improving the performance of functions that access surfaces.

Wednesday

  • Finished performance improvements: average_color is 4x faster, threshold is 2x faster, and mask.from_surface is 2x faster when using per pixel alphas.
  • Documented connected_component a little better

Thursday

  • Added support for a second surface back into threshold
  • Applied the same performance improvements to average_surfaces
  • Got rid of the gcc warnings that I could get rid of

Friday

  • First attempts at integrating my camera stuff with Physics
  • Wrote an example script that shows the basic process of combining get_image, colorspace, average_color, threshold, connected_component, and centroid to do hue based object tracking.