Liver CT image segmentation method and device based on deep learning neural network

The invention discloses a liver CT image segmentation method and device based on a deep learning neural network. The method comprises the following steps that: constructing a U-shaped DenseNet two-dimensional 2D network and a U-shaped DenseNet three-dimensional 3D network based on a dense convolutional network DenseNet and a U-shaped network Unet; integrating the U-shaped DenseNet 3D network intothe U-shaped DenseNet 2D network based on an auto-context method, obtaining a U-shaped DenseNet hybrid network, and performing deep learning training; and segmenting a liver CT image through the trained U-shaped DenseNet hybrid network. According to the method, the intra-layer information and inter-layer information of the CT liver image can be extracted as much as possible, meanwhile, calculationperformance is considered, and operation parameters are reduced, so that a computer can quickly output a relatively accurate liver and tumor segmentation image.