Text this: SF-SAM-Adapter: SAM-based segmentation model integrates prior knowledge for gaze image reflection noise removal