Tadas Baltrusaitis
452d1b8008
Integrating MTCNN in all of the demos.
2017-08-18 20:57:19 +01:00
Tadas Baltrusaitis
a251415fea
Starting to incorporate MTCNN
2017-08-18 10:03:01 +01:00
Tadas Baltrusaitis
441777c9ed
Code cleanup and rearrangement.
2017-08-17 12:35:11 +01:00
Tadas Baltrusaitis
8d01ea5911
Different implementation of convolution, through matrix multiplication which is quite a bit faster.
2017-08-17 11:40:57 +01:00
Tadas Baltrusaitis
2bd085cb47
Exploring CNN optimization
2017-08-16 17:14:00 +01:00
Tadas Baltrusaitis
d062ab690d
Final bit of simplification and fixing an issue with RNet and ONet.
2017-08-15 21:53:25 +01:00
Tadas Baltrusaitis
abcc91364e
Towards MTCNN, implementing PNet
2017-08-10 15:30:35 -04:00
Tadas Baltrusaitis
13e42a70cf
Some more work on MTCNN.
2017-08-09 16:16:31 -04:00
Tadas Baltrusaitis
7fe7bb4904
Towards MTCNN c++ implementation.
2017-08-09 15:04:14 -04:00
Tadas Baltrusaitis
8dea625717
Adding more preparation for MTCNN in C++
2017-08-09 12:01:53 -04:00
Tadas Baltrusaitis
8b5f8b9a36
Starting to work on MTCNN C++ implementation.
2017-08-09 11:00:38 -04:00
Tadas Baltrusaitis
5959176921
Some big changes:
...
- face validator is a retrained CNN now
- starting retiring CLM-Z from OpenFace
2017-08-01 17:11:02 -04:00
Tadas Baltrusaitis
8eb086545c
Simplified validator model.
2017-07-31 15:55:29 -04:00
Tadas Baltrusaitis
7513cf7964
Speeding up fancy face validation.
2017-07-29 21:11:16 -04:00
Tadas Baltrusaitis
8cc98be724
First try at a new landmark validator.
2017-07-29 18:19:41 -04:00
Tadas Baltrusaitis
305af01326
Auto stash before merge of "develop" and "origin/develop"
...
- Adding new license files
- Replacing images with more suitable CC ones
2017-05-08 21:36:23 -04:00
somacoder
86341fae30
Update LandmarkDetectorModel.h
...
Moving default boolean eye model assignment of false to cpp.
2017-02-23 13:45:09 -05:00
Guillaume Chanel
358cd6f9e2
Added the possibility to change the output videos codec by specifying the -oc option
...
A warning is now issued in the case the VideoWriter could not be instantiated, advising to change the codec
This change applies to FaceLandmarkVid* and FeatureExtraction executables
The option is used as "-oc CODEC_FOURCC" and defaults to DIVX. If the user inputs a FOURCC with a size different than 4, than the default codec is used
The option is handled by the function get_video_input_output_params in the LandMarkDetector namespace
2016-09-14 09:00:57 +02:00
Tadas Baltrusaitis
c1ff40399e
Bug fixes
2016-09-06 12:32:33 -04:00
Tadas Baltrusaitis
c1431fe815
Fixes with drawing antialiased lines and landmarks using sub-pixel precision.
2016-09-06 10:20:31 -04:00
Tadas Baltrusaitis
51ae195124
A number of small fixes, renamings and corrections.
2016-07-31 16:31:17 -04:00
unknown
57e58a6949
Master commit of OpenFace.
2016-04-28 15:40:36 -04:00
