Video Instance Segmentation with YOLOv8

This project implements video instance segmentation using YOLOv8, allowing real-time object detection, segmentation, and tracking in video streams. The notebook includes data preparation, training, inference, and model export for deployment.

Features

• YOLOv8-based instance segmentation for real-time multi-object tracking.
• Google Colab & Drive integration for cloud-based training and dataset management.
• ONNX model export for optimized inference and deployment.
• Custom dataset support with flexible preprocessing options.

Requirements

• Python 3.8+
• Google Colab or a local machine with GPU support.
• Installed dependencies (see Dependencies Imported in the Notebook).

Dataset Preparation

1. Upload your dataset to Google Drive and mount it in the notebook.
2. Ensure your dataset is in the YOLO format (images and annotation files).
3. Update the dataset path in the notebook before training.

Usage

1. Clone this repository or download the notebook.
2. Open the notebook in Google Colab or Jupyter Notebook.
3. Follow the step-by-step instructions to train and evaluate the model.

Running the Notebook

• Run all cells sequentially to set up the environment, train the model, and perform inference.
• For local execution, ensure you have all required dependencies installed.

Dependencies Imported in the Notebook

• torch, torchvision (Deep Learning framework)
• opencv-python (Image processing)
• numpy (Matrix operations)
• ultralytics (YOLOv8 framework)
• onnxruntime (ONNX model inference)

Results

• The trained YOLOv8 model can perform real-time video instance segmentation, detecting and tracking multiple objects simultaneously.
• The model can be exported to ONNX for optimized inference and deployment.

For further enhancements, consider fine-tuning hyperparameters, improving dataset quality, and deploying the model via FastAPI.