PUPIL-Detection-using-OpenCV

PUPIL detection from eyes images using OpenCV 3 and Python 3.6.

Face detection and Eyes detection using OpenCV is a common project which can be demonstrated using HAARCASCADEs by OpenCV. OpenCV have provided various trainers which can be used directly in detecting Faces, Eyes and other various elements.

PUPIL detection

Detecting PUPIL or EyeBall using OpenCV.

Algorithm

1. First take the eye image.
2. Make it invert.
3. Convert it to gray scale.
4. Apply Erosion Transform.
5. Use binary filter taking threshold value 220.
6. Find the biggest object.
7. Find that object's center point and height.
8. Highlight that circle.

detect_pupil_v2.py

Color Image Denoising and Gaussian Blur techniques are used for pre-processing the eye image before applying inversion filter over it. For segmentation, the contour with center nearest to center of image is filtered. Eyeball will be near to center of eye image Rest algorithm for pupil detection is same as in detect_pupil.py

Useful Links

Morphological Erosion : https://docs.opencv.org/3.0-beta/doc/py_tutorials/py_imgproc/py_morphological_ops/py_morphological_ops.html

OpenCV contours and Hierarchy : https://github.com/eyantrainternship/eYSIP_2015_Marker_based_Robot_Localisation/wiki/Contours-and-Hierarchy

Denoising Images : https://docs.opencv.org/3.2.0/d5/d69/tutorial_py_non_local_means.html

Smoothing Images : https://docs.opencv.org/3.1.0/d4/d13/tutorial_py_filtering.html