超音波影像分類系統及其方法

Disclosed is an ultrasonic image classification system and method thereof. The classification system includes: an ultrasound module configured to generate one or more ultrasound signals; a synthetic mastoid image module configured to convert the one or more ultrasound signals into one or more synthetic signals mastoid image; a synthetic mastoid image augmentation module configured to generate one or more augmented images composed of the one or more synthetic mastoid images through an imaging technique that combines different slices of the one or more synthetic mastoid images; a training module including a pre-trained deep learning model that is configured to receive the one or more augmented images and train the deep learning model accordingly; and a classification module configured to generate one or more classification results corresponding to the one or more augmented images according to a discriminant function obtained or continuously revised by the deep learning model.