Describir: Development and validation of a cranial ultrasound imaging-based deep learning model for periventricular-intraventricular haemorrhage detection and grading: a two-centre study.