CystNet: An AI driven model for PCOS detection using multilevel thresholding of ultrasound images

Area of Science:

• Endocrinology and Medical Imaging
• Artificial Intelligence in Healthcare

Background:

• Polycystic Ovary Syndrome (PCOS) is a common endocrine disorder affecting women of reproductive age, characterized by excess androgens and ovarian follicle abnormalities.
• Current diagnostic methods for PCOS, relying on manual ultrasound image interpretation, are time-consuming and prone to errors.
• Automated systems are needed to streamline PCOS diagnosis and improve accuracy.

Purpose of the Study:

• To develop and evaluate an advanced automated system for detecting and classifying Polycystic Ovary Syndrome (PCOS) using ultrasound images.
• To enhance diagnostic accuracy and efficiency through AI-driven analysis of ovarian follicle characteristics.

Main Methods:

• Image pre-processing techniques including resizing, normalization, augmentation, Watershed, and multilevel thresholding for precise segmentation.
• Feature extraction using the proposed CystNet technique.
• PCOS classification via fully connected layers with 5-fold cross-validation and ensemble machine learning classifiers.

Main Results:

• The AI system achieved a commendable accuracy of [Formula: see text] using a fully connected classification layer with 5-fold cross-validation.
• An accuracy of [Formula: see text] was attained when employing an ensemble machine learning classifier.
• The system demonstrated robust performance across various evaluation metrics, including AUC score, precision, recall, and F1-score.

Conclusions:

• The proposed AI-based automated system effectively detects and classifies PCOS from ultrasound images, offering a more accurate and efficient diagnostic tool.
• This approach holds potential for predicting PCOS and similar conditions using multimodal datasets.
• The system can facilitate timely intervention and reduce the diagnostic burden on healthcare professionals.