Skip to main content
. 2023 Jan 11;23:20. doi: 10.1186/s12884-022-05204-x

Fig. 2.

Fig. 2

Workflow overview of the entire protocol. US videos are separated into frames, from which human experts assign the key planes to the four different categories. Data are anonymized, non-important frames are added as a distinct category of images, the model is trained on the obtained image data set and will then classify frames in new US videos and generate reports