| Commit | Date | |
|---|---|---|
| 2003-09-21 04:08:09 | Tree | |
|
[r493]
by
ltilib
wickel: fixed a bug in the prediction ellipse drawing code (radius of the |
2003-09-21 04:07:04 | Tree |
| 2003-09-21 04:05:43 | Tree | |
| 2003-09-21 04:05:03 | Tree | |
|
[r490]
by
ltilib
wickel: fixed bug: an image with a palette containing only RGB was not properly converted to lti::image format |
2003-09-21 04:04:23 | Tree |
| 2003-09-17 04:02:02 | Tree | |
|
[r488]
by
ltilib
alvarado: @FIX: An error in lti::watershedSegmentation has been fixed that forced the wrong flooding of valleys, merging too many regions. It was produced by the wrong neighbor validation check at the borders of the image. |
2003-09-15 10:04:16 | Tree |
|
[r487]
by
ltilib
alvarado: @ADD: New apply added to lti::computePalette, which computes the mean color from three channels and the number of pixels per region. This method existed already for lti::image but not for three lti::channel as input arguments. |
2003-09-15 10:04:05 | Tree |
|
[r486]
by
ltilib
alvarado: @CHG: The lti::watershedSegmentation used a very too old interface. Now, the apply methods return true or false as expected. Also the parameter "rainfall" has been changed to true, as the rainfall method is much faster and does not have the problem (or feature?) of wide watersheds. |
2003-09-15 10:03:52 | Tree |
| 2003-09-13 04:02:23 | Tree |