Figure 1.

Framework of the BETNET model. In the convolution layer: W = weight, b = bias, and BN = batch normalization. The maximum pool operation was used in the pool layer. Reshape is a process that changes the feature map with the size from the previous layer and is then input into the fully connected layer. Softmax mapped the output of the fully connected layer to the results of the classification. The white scale bars represent 1 cm on the ultrasound image.