Blender Git Statistics -> Branches -> libmv_prediction
"Libmv_prediction" branch
Total commits : 24
Total committers : 2
First Commit : September 26, 2014
Latest Commit : October 30, 2014
Commits by Month
| Date | Number of Commits | |
|---|---|---|
| October, 2014 | 18 | |
| September, 2014 | 6 | |
Committers
| Author | Number of Commits |
|---|---|
| Sergey Sharybin | 23 |
| Keir Mierle | 1 |
Popular Files
| Filename | Total Edits |
|---|---|
| tracking_auto.c | 7 |
| tracking_util.c | 6 |
| CMakeLists.txt | 4 |
| autotrack.cc | 4 |
| marker.h | 4 |
| tracksN.cc | 3 |
| tracksN.h | 3 |
| bundle.sh | 2 |
| autotrack.cc | 2 |
| frame_accessor.cc | 2 |
Latest commits 
October 30, 2014, 18:29 (GMT) |
Libmv: Fix for famous Image(x,y,z) vs. Image(y,x,z) |
October 30, 2014, 18:05 (GMT) |
Merge branch 'master' into libmv_prediction |
October 30, 2014, 17:38 (GMT) |
Libmv: Code cleanup |
October 30, 2014, 17:25 (GMT) |
Libmv: Inverse meaning of Marker::visible_channels This way it's not needed to initialize channels bits if all of them re expected to be used, plus simplifies checks in some cases as well. |
October 30, 2014, 17:05 (GMT) |
Merge remote-tracking branch 'origin/master' into libmv_prediction |
October 30, 2014, 17:03 (GMT) |
Libmv: Support disabled color channels in trackign settings This was never ported to a new trackign pipeline and now it's done using FrameAccessor::Transform routines. Quite striaghtforward, but i've changed order of grayscale conversion in blender side with call of transform callback. This way it's much easier to perform rescaling in libmv side. |
October 29, 2014, 11:52 (GMT) |
Merge branch 'master' into libmv_prediction |
October 29, 2014, 11:46 (GMT) |
Libmv: Fix compilation error with libmv disabled |
October 24, 2014, 13:40 (GMT) |
Libmv: Fix read/write pass the array end |
October 23, 2014, 09:39 (GMT) |
Libmv: Fix wrong search area clamping in frame accessor Not sure who was that smart to make it wrong :) Hopefully it'll fix the "magnetic" borders reported by Sebastian, when the marker gets "attracted" by the frame bounds. |
October 23, 2014, 08:55 (GMT) |
Merge branch 'master' into libmv_prediction |
October 20, 2014, 21:47 (GMT) |
Libmv: Fix stupidness in termination criteria Need to skip current track, not to stop syncing current frame. |
October 20, 2014, 21:23 (GMT) |
Merge branch 'master' into libmv_prediction |
October 20, 2014, 21:10 (GMT) |
Libmv: Correction for previous commit - wrong axis used for AABB calculation |
October 20, 2014, 18:41 (GMT) |
Libmv: Bring back track margin check |
October 20, 2014, 18:18 (GMT) |
Libmv: Use proper termination criteria based on the region tracker result |
October 20, 2014, 17:54 (GMT) |
Fix compilation error with disabled libmv |
October 20, 2014, 17:53 (GMT) |
Merge branch 'master' into libmv_prediction |
September 27, 2014, 14:15 (GMT) |
Tracking: Fix crashes and some spontaneous track loss Crash was caused by the fact that requested region might go outside of the actual frame. For now we deal with this by only filling the region partially with the data we can and keep rest of the image empty. There also used to be a bug in the detection that tracking failed caused by reading uninitialized variable. |
September 27, 2014, 09:49 (GMT) |
Libmv: Correct include dir to PNG library |
MiikaHweb - Blender Git Statistics v1.06
MiikaHweb | 2003-2021