Abu Anas Shuvom
Abu Anas Shuvom
You can train another class named "Unknown face" where you can add faces of different persons and then train again. Thus the network will learn to differentiate unknown faces.
This might be an issue with your OpenCV installation or Gstreamer either. I see you are using Gstreamer for video capture although the command takes OpenCV video capture functions. Could...
With adequate confidence level (threshold), the DNN simply won't detect the face, let alone recognition. Can u try this to validate?
The cfg file doesn't seem to be 30 layers (or 15 layers for tiny) . Check the .cfg file, it should be like this `[net] batch=1 subdivisions=1 width=416 height=416 channels=3...
What is your Batch size and sub-divisions? Reduce the numbers according to your GPU capacity. Also you may try with` OPENCV=0` for training.
Which GPU are you using?
@Isabek Hi, did you trained ur own model with lower class of objects? Does it improve fps?
Can't encode transcription: 'ঢাকা মেটো-গ' in language '' Encoding of string failed! Failure bytes: ffffffe0 ffffffa6 ffffffbe ffffffe0 ffffffa6 ffffff95 ffffffe0 ffffffa6 ffffffbe 20 ffffffe0 ffffffa6 ffffffae ffffffe0 ffffffa7 ffffff87...