Spatially Informed Analysis of Ovarian Ultrasound Images for Improved PCOS Diagnosis
Year: 2026
Type: conference-paper
Abstract: Polycystic Ovary Syndrome (PCOS) is usually diagnosed from ovarian ultrasound images based on follicle count. However, follicle count alone does not describe how follicles are spatially arranged within ovaries. This paper proposes an automated PCOS analysis framework that focuses on follicle spatial organization by combining an improved Weighted Hausdorff Distance (WHD) with radial distribution entropy. The method was evaluated on 6,784 ovarian ultrasound images, all processed successfully. The results show that 81.2%of cases display PCOSlike patterns, characterized by high follicle density and organized spatial structure. Both statistical analysis and machine learning experiments identify WHD as the most influential feature for PCOS characterization, indicating that spatial geometry provides more reliable diagnostic insight than follicle count alone.
Authors: Sulagna Mondal, Ishani Das, Tuli Bakshi
Language: English
Subfield: Reproductive Medicine
Field: Medicine
Domain: Health Sciences
Open Access status: hybrid
OpenAlex ID: W7203621132