Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Detection is glitchy and buggy #12

Open
Schattenkobold opened this issue Apr 9, 2021 · 0 comments
Open

Detection is glitchy and buggy #12

Schattenkobold opened this issue Apr 9, 2021 · 0 comments

Comments

@Schattenkobold
Copy link

Hey,

I'm currently trying to make this work correctly and I'm having trouble with making it look anywhere nearly as good as what's shown in the demo video.

I tried using the programm on 2 different PCs and tested severall python, opencv and dlib versions including the ones which you list in the readme. I'm using Unity-Version 2018.4.33f1 for every try. Everytime the following happens when the python window opens and I see myself + the rotation detection marks:

In case of the CubeHanHan Model:

  • The whole head just disappears and is not seen in the game window of unity
    In case of theKizunaAI Model:
  • eye pupils disappear and eyes are very open even when the debug for the eye detection shows that it detects the eyes to be closed (The graph line is always at the bottom of the green debug window on the canvas)
  • the head glitches around seemingly randomly. I think it also wants to dissappear like the CubeHanHan Model but can't because it's attached to the body. The rotation seems to work okayish. It at least reacts somewhat to me spinning my head.

I found no info to problems like this in the readme or other Issues. Here is something I did which makes the glitches a bit less:

In ParameterServer.cs I found 2 blocks of code on the line 155 to 163. I commented the lines which look like this:

getheadRot.w = W_Kalman.kalman_filter(getheadRot.w, 8e-3f,5e-4f);

and uncommented the lines which look like this:

getheadRot.w = W_Kalman.kalman_filter(getheadRot.w, 1e-4f,1e-4f);

Did you experience anything like the problem I'm describing during development? I'm sadly too much of an beginner with facial detection, opencv and dlib to understand where things go wrong with the detection and/or the applying of the filters to figure it out myself.

Yours sincerely,
S.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant