Assisting Ultrasound Examination Based on Transfer Learning Method

Ultrasound examination is the preferred method for related diseases. The transfer learning method is considered to be a feasible and automatic method for optimizing the ultrasound examination. This paper uses the transfer learning method the diagnosis of hepatic hemanguimas. First, establish an optimized deep learning technology model. A stacking convolutional self-coded depth model is designed to deal with the lack of label image dataset Based on CNN. Second, External transfer learning technology model is established to the automatic classification diagnosis of ultrasound images according to the transfer learning technique. Comparing with the two results and combined with the artificial examination in our experiments, the diagnosis based on the transfer learning method can achieve an acceptable result.