facebookresearch supervision-by-registration issues

facebookresearch / supervision-by-registration

Supervision-by-Registration: An Unsupervised Approach to Improve the Precision of Facial Landmark Detectors

Other

760 stars 165 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

.dat file

#80 jayp84847 opened 3 years ago
0
Where is the snapshot

#79 dolphins22365 opened 3 years ago
0
300VW Training

#78 yongyongdown opened 3 years ago
0
Training on another landmark problems

#77 eanemo opened 3 years ago
0
Torch Script

#76 eusebioaguilera closed 4 years ago
0
Where pam is applied

#75 makalo opened 4 years ago
2
How to debug LK Module

#74 sljlp opened 4 years ago
3
The steps

#73 sevaraMS opened 4 years ago
0
Face Mesh Results？

#72 CodingMice opened 4 years ago
5
Pretrained Model

#71 ShakedDovrat closed 4 years ago
1
test on own video

#70 fnhdx opened 4 years ago
0
Result for pretrained model is very bad and unstable at angled face. Is the pretrained model that bad?

#69 yxchng closed 4 years ago
9
About using camera

#68 come880412 closed 4 years ago
0
Problem about training on a small trainset

#67 JINXvvv opened 4 years ago
1
Training problem in the first stage

#66 KeyKy closed 4 years ago
4
what's the meaning of meta files?

#65 JINXvvv closed 4 years ago
2
Where to get the pretrained model?

#64 fanyhchn closed 4 years ago
2
Base Detector?

#63 ModarD closed 4 years ago
1
Train on my viedo dataset

#62 HansonXia opened 4 years ago
11
What is the size of the training model?

#61 hyb1234hi closed 4 years ago
6
vis.py

#60 Waxyoung closed 5 years ago
0
The effect of landmark detection dramatically decreases as the distance increases

#59 xfj81525 opened 5 years ago
3
replace the backbone

#58 AaronXinWu opened 5 years ago
6
About how get the landmark of Reg

#57 xyzhai-pro closed 3 years ago
5
feature extraction with MobileNet hardly converge

#56 xfj81525 opened 5 years ago
0
Train my own dataset

#55 abdou31 closed 4 years ago
8
the LK operation is very slow, is this normal?

#54 taozhiqi closed 4 years ago
1
maybe something wrong in transforms.py

#53 momiji7 closed 4 years ago
2
demo-detsbr-vis.mp4 and demo-sbr-vis.mp4 have almost no differences?

#52 CodingMice closed 4 years ago
1
evaluate a single image

#51 JiHaoqina closed 5 years ago
4
To improve efficiency, you can change the backbone of our SBR, which is VGG-16 (see https://github.com/facebookresearch/supervision-by-registration/blob/master/lib/models/cpm_vgg16.py). Besides, we use an input size of 224, you can downsample the input to 128, or 96, or 64.

#50 xfj81525 closed 5 years ago
0
About the optimization of LK

#49 zhang-zx opened 5 years ago
16
Doubts about Video Tags

#48 zrqi closed 5 years ago
4
About face bounding box

#47 pokidyshev closed 5 years ago
12
About speeding up the LK

#46 zhang-zx closed 5 years ago
1
confused about the NMS

#45 wqz960 closed 4 years ago
5
evaluate the 300W video

#44 cploveflower closed 5 years ago
1
How to make the model smaller?

#43 zhang-zx closed 5 years ago
5
Problems when training on 300W

#42 zhang-zx closed 5 years ago
6
How to get the coordinates from the heatmaps？

#41 SunnyDp closed 5 years ago
1
Can I use my own video to test the cpm-sbr?

#40 dyther closed 5 years ago
3
Reproducing quality of detection comparable with demo.gif

#39 Daiver closed 5 years ago
4
can this project be transferred to mobile services

#38 imerse closed 5 years ago
1
The loss between 'batch_location' and the groundtruth

#37 SunnyDp closed 5 years ago
1
Training landmark detectors using my own dataset such as 300W-LP

#36 SunnyDp closed 5 years ago
1
Some questions about the different configs in the "specify_parameter" function in the cpm_vgg16.py

#35 SunnyDp closed 5 years ago
1
landmarks still shake after 10 epochs

#34 Racha1992 closed 5 years ago
12
Regarding to lk_config.weight, lk_config.stable and so on.

#33 taey16 closed 5 years ago
1
Some questions about preparing data

#32 tongxinjie closed 5 years ago
1
Error for eval.py

#31 NikolayTV closed 5 years ago
3