the model yolov8sbest.pt
is trained using roboflow
and colab
and used here for inference.
the streamlit
application is launched on the local server and shows available cameras for surviellance.
activate virtual environment with the said dependencies and then run the webapp by streamlit run app.py
in the main directory.
each camera calls the process
method from the base
file and will save images with confidence>0.6 and 5 continuous frames in user directory.