Multiple object detection

Prerequisites

This application was developed and tested on a Linux/WSL2 environment. Functionality on other operating systems has not been tested and is not guaranteed.

1. Operating System: Linux / WSL2
2. Python 3.10.12: We recommend using pyenv for managing Python versions. If pyenv is not installed, you can install it following the instructions here. After installing pyenv, you can install Python 3.10.12 using pyenv install 3.10.12.
3. Recomended enviroment manager is pipenv, but its not mandatory.
4. Follow this instructions

Try detection with app

Installation and usage

1. Clone the repository
2. Navigate to the project directory
3. Set the local Python version to 3.10.12
4. Create and activate new virtual environment
5. Install the required dependencies: pipenv install or pip install -r requirements.txt
6. To run the application for video processing: streamlit run app.py

Try train your own model

At first, you need have collected some dataset. How should looks annotations, check official docs from ultralytics. Images do not have to be in the same resolution, or format. The only thing you need to pay attention to is that a file with annotations (dog.txt) is available in the same folder for each image (dog.png)

Dataset

1. Create in root directory this structure: data/dataset
2. Inside of dataset directory add images (supported formats are jpg and png) and anotations (txt) files
3. Into data directory insert obj.names file
4. Be sure that you can run Jupyter Notebooks on your machine
5. Open data_processing.ipynb and run it.
6. Script will create some new directories in data directory and will do some magic there

Training

1. Be sure that your machine is supported and you have installed CUDA 12.1.
2. Run train.py script
3. Go to sleep, or make coffee bcs this will take some time....
4. If training will be success, new model will be created and sreamlit app will find it and will use this model as default.

Contributing

Contributions are welcome! Please feel free to submit a pull request.

Setting Up in Visual Studio Code

If you are using Visual Studio Code as your IDE, you may need to specify the path to the Python interpreter in the .code-workspace file..

Here's how you can do it:

1. Open the .code-workspace file in your workspace.
2. Look for the settings object
3. Inside the settings object, add or modify the python.defaultInterpreterPath property. The value should be the absolute path to the Python interpreter you want to use.

License

This project is licensed under the MIT License.