Method of automatic segmentation of maxillofacial bone in CT image using deep learning

A non-transitory computer-readable storage medium having stored thereon program instructions, the program instructions executable by at least one hardware processor to:

• receive input CT slices of a CT image including a maxillofacial bone;
• segment the input CT slices into a mandible and a portion of the maxillofacial bone excluding the mandible using a convolutional neural network structure, the convolutional neural network structure including an encoder including a first operation and a second operation different from the first operation in a same layer and a decoder including a third operation and a fourth operation different from the third operation in a same layer;
• and
• accumulate 2D segmentation results which are outputs of the convolutional neural network structure to reconstruct a 3D segmentation result,
• wherein the encoder includes a first skip connection adding a previous feature map to a present feature map in the same layer of the encoder,
• wherein a size of an input of a convolutional operation immediately after the first skip connection of the encoder is same as a size of an output of the convolutional operation immediately after the first skip connection of the encoder, and
• wherein a number of channels of the outputs of the convolutional operation immediately after the first skip connection of the encoder is twice a number of channels of inputs of the convolutional operation immediately after the first skip connection of the encoder.