This repository contains a Python script that processes a video to detect and highlight the region of interest (ROI) around the eye. The script performs real-time video analysis, applying various image processing techniques to isolate and track eye movement.
Grayscale Conversion:Converts the region of interest (ROI) to grayscale for easier processing.Gaussian Blur:Applies a Gaussian blur to reduce noise and smooth the image.Thresholding:Uses thresholding to create a binary image, isolating the eye for contour detection.Contour Detection:Identifies and highlights the largest contour (presumed to be the eye) with bounding boxes and crosshairs.Real-time Visualization:Displays the processed video frames in real-time.
To run this script, you need to have Python installed along with the following dependencies:
pip install opencv-python numpy- Clone the repository:
git clone https://github.com/akhilaprabodha/Eye-Motion-Tracking.git
cd Eye-Motion-Tracking- Run the script:
python eye_motion_tracking.py- Script Explanation:
- The script opens a video file (eye_recording.flv) or captures video from a camera.
- It defines the region of interest (ROI) within the video frame to focus on the eye.
- The ROI is processed through grayscale conversion, Gaussian blur, and thresholding.
- Contours are detected and the largest one is highlighted with a bounding box and crosshairs.
- The processed frames are displayed and saved as three separate output videos.
1. Thresholded Output 🔲
- This video shows the thresholded version of the ROI, where the eye is isolated in a binary image.
2. Grayscale ROI Output 🌑
- This video displays the grayscale version of the ROI, which is used as the base for further processing.
3. Contour Detection ROI Output 🎯
- This video highlights the ROI with the detected contour, bounding box, and crosshairs.
Contributions are welcome! If you have any improvements or additional features in mind, feel free to open a pull request.