opentrack.git/tracker-points, branch feature_windows_dark_theme head tracking software for Windows and Linux http://ananke.misaki.pl/cgit/opentrack.git/atom?h=feature_windows_dark_theme 2019-04-24T16:46:12Z Renaming Points Tracker to Easy Tracker. 2019-04-24T16:46:12Z Stéphane Lenclud github@lenclud.com 2019-04-01T18:31:10Z urn:sha1:c029e52a330842415502cc29a3460e016d4a8a93 Easy Tracker: Preview cross hair now showing top point. 2019-04-24T16:46:12Z Stéphane Lenclud github@lenclud.com 2019-04-01T18:18:02Z urn:sha1:456a922b5995f1f836c13c5795258bc83e521571 Deleting home made P3P solver we branched from pt now that our OpenCV solution is working. Translations now in centimetres. 2019-04-24T16:46:12Z Stéphane Lenclud github@lenclud.com 2019-03-31T20:01:36Z urn:sha1:eb77f73826a26ad1dea84e3ea0173bb952a46e26 Tracker now providing OpenCV data. 2019-04-24T16:46:12Z Stéphane Lenclud github@lenclud.com 2019-03-31T19:39:04Z urn:sha1:c7df551561ca7bd7ae22abecc562cd882259f748 Fixing look-up of top point after realizing that our bitmap origin is top left. 2019-04-24T16:46:12Z Stéphane Lenclud github@lenclud.com 2019-03-31T18:50:58Z urn:sha1:af000c404a003c046635890863f485b159ec7107 Fixing cap clip model. Using OpenCV APIs to get Euler angles, though it's possibly slower as it is doing a bunch of other things too. It looks like angles are correct except that we are getting some yaw when pitching down. Adding Euler angles computation and distortion coefficient. 2019-04-24T16:46:12Z Stéphane Lenclud github@lenclud.com 2019-03-31T08:24:52Z urn:sha1:2c88602928c20c1bea05fc15179c0c4866b42f8b Angles values are being funny though, looks like it will take some serious debugging to get something that makes sense. Though I guess from there we should try feeding the OpenCV computed value to OpenTrack and trash the original pt tracker code. First solveP3P results that are looking consistent. 2019-04-24T16:46:12Z Stéphane Lenclud github@lenclud.com 2019-03-31T07:40:37Z urn:sha1:feb7026316a4f2ad551b4ea87226c264c5277ca4 Translation vector in meters seems to be spot on. Rotation angles still need to be computed. Radial distortion still need to be taken into account. Hacking my way through to cv::solveP3P usage. 2019-04-24T16:46:12Z Stéphane Lenclud github@lenclud.com 2019-03-28T19:54:35Z urn:sha1:8141c4f07b1ddc4555d10a78ea5c3f482c8be04f Create Points Tracker based on original Point Tracker. 2019-04-24T16:46:12Z Stéphane Lenclud github@lenclud.com 2019-03-27T20:18:39Z urn:sha1:b0e22e955d566b2a981cb43ae477f2646c0c978d