search

justusschock / shapenet

PyTorch implementation of "Super-Realtime Facial Landmark Detection and Shape Fitting by Deep Regression of Shape Model Parameters" predicting facial landmarks with up to 400 FPS
https://shapenet.rtfd.io
GNU Affero General Public License v3.0
341 stars 59 forks source link

problem with prediction #31

Open satish1901 opened 4 years ago

satish1901 commented 4 years ago

Does the code always a need a "landmark.pts" files to detect the final face? why can't i just detect the face using a direct input image? Can you please clarify if i am doing something wrong here?

justusschock commented 4 years ago

You can do this, but the prediction script was written to calculate precision and accuracy scores. Therefore you always need a ground truth. You can, however, copy the script and just remove the parts that require a ground truth

satish1901 commented 4 years ago

Thank you Can you please share the trained weights on Thermal images?

justusschock commented 4 years ago

Hi, Unfortunately I cannot do this (since for thermal images I only got weights for a more advanced, internal but not yet published, method) . But it has been trained on the database described in https://github.com/marcinkopaczka/thermalfaceproject . So if you write my colleague there a mail, you should easily be able to train it yourself :)