Unable to link app with Unity

[tulaibpirzada]

Hello,
I am facing wierd issue. The offline mode works fine but when I try to run it real time, it doesn't seem to connect with openface app. There is error showing anywhere whatsoever, still not working. Need some guidance.

Thanks in advance

[NumesSanguis]

Hi,

It would be hard to identify the problem with just this information. Could you tell me the following (and any other relevant) information?

• What OS are you working on?
• What steps did you take, and where does it fail? For example, the Python modules are passing on the data, but it is not received by Unity3D.
• What versions of software are you using? The modified OpenFace which includes ZeroMQ (https://github.com/NumesSanguis/FACSvatar/releases), or did you get it from the original OpenFace GitHub (https://github.com/TadasBaltrusaitis/OpenFace/)? What Unity3D version.
• Any screenshots.

[tulaibpirzada]

Hello, following are my replies:

1. Windows 10 Home
2. Ok so it working fine until Offline version. The python code that we run make the character face move in Unity. But when I try do real time webcam detection in Unity, it doesn't work.
3. Yes I am using the latest release.
4. Unity version is 2018.3.14f1
5. There is no error coming, what screenshots you are requiring?

[NumesSanguis]

Just to make sure, are you using openface_2.1.0_zeromq.zip from this link?: https://github.com/NumesSanguis/FACSvatar/releases and did you follow the step 0. Downloads under Quickstart (NEW)?

Are you using the FACSvatar modules with Docker or just by using a Python environment on your computer and then running the modules separately?

For screenshots I mean the Python terminal. If the OpenFace application is correctly sending the data, the process_bridge module (FACSvatar\modules\process_bridge) should receive data and some output should be generated in your terminal. Something like this screenshot:

That image was created with the following steps:

1. Make sure that inside the openface_2.1.0_zeromq\config.xml the IP is: <IP>127.0.0.1</IP> (if run on the same computer, otherwise change this IP).
2. Start OpenFace by double clicking OpenFaceOffline.exe
3. (Non-Docker approach) Navigate with a terminal to FACSvatar\modules\process_bridge and run the command: python main.py
4. You should see output in the terminal appear.

[uumarfarooq]

Hello,
I am facing same issue did all steps.
offline mode is working fine.

Please Help me

Thanks

[tulaibpirzada]

Just to make sure, are you using openface_2.1.0_zeromq.zip from this link?: https://github.com/NumesSanguis/FACSvatar/releases and did you follow the step 0. Downloads under Quickstart (NEW)?
----> Yes I am using same link
Are you using the FACSvatar modules with Docker or just by using a Python environment on your computer and then running the modules separately?
----> I am using docker.
For screenshots I mean the Python terminal. If the OpenFace application is correctly sending the data, the process_bridge module (FACSvatar\modules\process_bridge) should receive data and some output should be generated in your terminal. Something like this screenshot:

That image was created with the following steps:

1. Make sure that inside the openface_2.1.0_zeromq\config.xml the IP is: <IP>127.0.0.1</IP> (if run on the same computer, otherwise change this IP).
   ----> Yes I am running on same system using above mentioned IP
2. Start OpenFace by double clicking OpenFaceOffline.exe
   ----> Already doing
3. (Non-Docker approach) Navigate with a terminal to FACSvatar\modules\process_bridge and run the command: python main.py
   ----> I am using docker approach
4. You should see output in the terminal appear.
   ---> Outputs are coming but Unity face is not detecting real time

[NumesSanguis]

I only now realize that you're using Windows 10 Home, sorry.
For that reason, you cannot use your local ip. Instead you have to use the Docker container's IP.
Therefore, step 1. was wrong.

Please see these instructions: https://github.com/NumesSanguis/FACSvatar#with-webcam-for-real-time-windows-only-for-now

---

Could you already solve your issue? Did you take the Python instead of Docker approach?