Anomaly detection in optical networks is vital for network reliability. We investigate leveraging state of polarization (SOP) data and computer vision for this task, despite challenges like scarce anomaly data. Our study explores using deep learning to identify anomalies in optical networks, focu…