Eye Tracking System

This repository contains the code for a basic eye tracking system that processes video frames to detect and highlight a region of interest (ROI) around the eye. The system detects contours in the specified ROI, identifies the largest contour, and draws a bounding rectangle and crosshairs on it to indicate the position of the eye.

Features

• Real-time Processing: Processes video frames in real-time to detect the eye's position.
• Region of Interest (ROI): Focuses on a specific area of the video frame to reduce computational load and improve accuracy.
• Contour Detection: Identifies the largest contour within the ROI and marks it with a bounding rectangle.
• Crosshairs: Draws crosshairs over the detected eye to visualize its center.

Prerequisites

Before running the code, ensure you have the following installed:

• Python 3.x
• OpenCV (cv2)
• NumPy (numpy)

You can install the required Python packages using pip:

pip install opencv-python numpy

Getting Started

1. Clone the repository:

   git clone https://github.com/Navini11/eye-tracking-system.git
   cd eye-tracking-system

2. Add your video file:

   Ensure that the video file (eye_recording.flv) is located in the same directory as the script or update the path in the code if it’s located elsewhere.

3. Run the code:

   python eye_tracking.py

   This will start processing the video and display the following windows:

   • Threshold: Shows the binary thresholded image used for contour detection.
   • Gray ROI: Displays the grayscale, blurred version of the ROI.
   • ROI: Shows the original ROI with the detected contours, bounding rectangle, and crosshairs.

4. Exit the program:

   The program will continue to process and display frames until you press the Esc key.

Code Explanation

Main Components

• Video Capture: The video is loaded using cv2.VideoCapture().
• Region of Interest (ROI): A specific part of the frame is defined as the ROI where the eye is expected to be.
• Grayscale and Gaussian Blur: The ROI is converted to grayscale and blurred to reduce noise.
• Thresholding: A binary threshold is applied to highlight the eye.
• Contour Detection: Contours are detected, and the largest one is assumed to be the eye.
• Drawing: A bounding rectangle and crosshairs are drawn around the detected contour.

Key Functions

• cv2.cvtColor: Converts the image to grayscale.
• cv2.GaussianBlur: Applies a Gaussian blur to the image.
• cv2.threshold: Applies a binary threshold to the image.
• cv2.findContours: Finds the contours in the thresholded image.
• cv2.boundingRect: Calculates the bounding rectangle for a contour.
• cv2.rectangle: Draws a rectangle on the image.
• cv2.line: Draws lines (crosshairs) on the image.
• cv2.imshow: Displays the image in a window.

Customization

• ROI Size: Modify the ROI coordinates (frame[269:795, 537:1416]) in the code to adjust the area being processed.
• Threshold Value: Change the threshold value (3 in cv2.threshold) for different lighting conditions.
• Contour Area: You can add more logic to handle multiple contours if needed.

License

This project is licensed under the MIT License. See the LICENSE file for details.