Method for acquiring high-resolution and high-precision segmentation result of CT (Computed Tomography) image

• Main Authors: DING YUGUO, QIAO TIAN, ZHAI FANGWEN
• Format: Patent
• Language: Chinese; English
• Published: 02.07.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The invention provides a method for obtaining a high-resolution and high-precision segmentation result of a CT image, and the method does not pursue that a neural network directly outputs a finer segmentation result, and on the contrary, obtains a rough contour feature map of a skeleton from a small-scale strong-semantic CT image feature map by using the powerful feature extraction capability of a convolutional network, and obtains a high-resolution and high-precision segmentation result of the CT image. And performing direct linear sampling on the feature map to be consistent with the original CT resolution to obtain a rough contour segmentation result, then combining the rough contour segmentation result with the feature map subjected to the first layer of convolution, and then connecting four convolution modules as a high-resolution segmentation result to be output. Compared with the prior art, the method has the advantages that on one hand, the output of the large-scale high-resolution CT image is effecti