fivkovic / clocksy

⏰ A machine-learning-based analog clock recognition and time detection app developed during the Soft Computing university course.

⏰ A machine-learning-based analog clock recognition and time detection app developed during the Soft Computing university course.

Requirements

• Python 3.7

Environment setup

1. Create a new virtual environment

   python3 -m venv myenv
   

2. Use requirements.txt to install all required dependencies.

   pip install -r requirements.txt
   

Running the application

Use

External stream

1. Download IP Webcam on your phone
2. Connect your PC and phone to the same local network
3. Start application and select Start Server option. The application will start capturing video and show you your IP address.
4. Use this IP address to run the controller.py script

python controller.py -s stream -su http://192.168.1.102:8080/video --width 720 --height 720

Webcam

python controller.py -s webcam --width 640 --height 480

Training

Run the scripts with default parameters.

Clock tracking

clock_tracking\core.py -m train

Time reading

time_reading\core.py -m train

Testing

Run the scripts with specifying the image count and noise threshold.

Clock tracking

clock_tracking\core.py -m test -ic 2000 -nt 0.5

Time reading

time_reading\core.py -m test -ic 2000 -nt 0.3

About the Team

|
_{Filip Ivković} |
_{Katarina Tukelić} | | --- | --- |