Abstract
Background The aim of this study is to use the computer aided diagnostic (CAD) technique to help doctors interpret and diagnose the ultrasound images of mammary glands. CAD technology can reduce doctors’ workload, improve objectivity and accuracy of ultrasound breast examination and help further diagnosis and treatment.
Methods This study chose to use the TensorFlow framework to train and predict the neural network. The GPU support version of TensorFlow’s Windows platform was selected, and the CUBA toolkit and cuDNN library were used to support the GPU operation.
Results Combined with the convolution neural network algorithm in the field of deep learning, an automatic segmentation method for breast ultrasound images is proposed, which transforms the image segmentation task into the classification task of each pixel in the image.
Conclusion The CAD method can correctly distinguish the tissue area of skin, gland and tumor in breast ultrasound images. The shape and contour are similar to the standard results manually labelled by doctors, and have achieved good results in various quantitative evaluation indexes. In the experiment, the segmentation results generated by different neural network parameter configurations were compared with other methods which showed that the proposed method in this study has certain advantages.