Machine Learning Method for Recognizing Human Eye Pupil Region

This research study describes how machine learning techniques can be employed to segment human pupils. The objective is to recognize and isolate the pupil region in ocular images. The aspects of a methodology are data set preparation, feature extraction, machine learning model selection, model training, model optimization, model testing, model evaluation, and finally post-processing. During preprocessing, this identifies and designates pupil regions in the dataset to obtain high-quality eye images. This compares two machine learning models, the Support Vector Machine (SVM) and the Random Forest, after identifying the most significant image characteristics. After being trained and fine-tuned using the dataset, models are evaluated based on metrics such as precision, recall, F1 score, and accuracy. In post-processing, the segmented pupil regions are modified after the initial segmentation. Random Forest attained 96.27 accuracy, while Support Vector Machine (SVM) obtained 95.17. This research contributes to the advancement of computer vision and is beneficial for software that employs human vision.