Automatic analysis method for cone beam CT image three-dimensional craniofacial structure

Abstract

The invention discloses an automatic analysis method for the cone beam CT image three-dimensional craniofacial structure. Based on a picture model and the full convolutional neural network, the picture model is trained for automatic detection and positioning of an anatomical structure, and the full convolutional neural network is utilized to construct mapping between a cone beam CT image of the anatomical structure and a corresponding marked image. At the test stage, for an input three-dimensional cone beam CT image, firstly, the picture model is utilized to detect the spatial position of theanatomical structure of interests, secondly, the full convolutional neural network is utilized to estimate automatic marking of image subblocks of the structure, and automatic three-dimensional craniofacial structure analysis is realized. The method is advantaged in that the three-dimensional craniofacial structure of interests in the cone beam CT image can be automatically segmented and marked, automatic analysis and segmentation of the stable structure are acquired, and the method can be applied to evaluation of marking and the curative effect of an oral orthodontic treatment plan.

Description

Technical field [0001] The invention relates to the field of computer vision and oral clinical medicine, in particular to a method for automatically analyzing three-dimensional craniofacial structures in cone beam CT images. Background technique [0002] The automatic analysis of craniofacial structure is the basis for clinical orthodontic treatment evaluation and surgical prediction. For example, overlapping cone-beam CT images before and after treatment to visualize and evaluate changes in craniofacial morphology caused by treatment and growth depends on cone-beam CT The locally stable structure of the segmentation in the image. Taking into account the local fine bone structure, it is still difficult to automatically segment the relatively stable structures in craniofacial orthodontics, such as the zygomatic arch and anterior skull base. Manual interactive segmentation is time-consuming and depends on the experience of relevant personnel. Cone-beam CT images have a relatively...

AI Technical Summary

Problems solved by technology

However, methods based on statistical surface or grayscale models are difficult to deal with fine structures in the reduced subspace

Annotation migration and fusion using image registration or template deformation methods often have high online computing costs

In recent years, convolutional neural networks have been used for feature detection and segmentation of medical images, but there is still a lack of effective technical solutions for end-to-end segmentation and labeling of small bone structures on large-scale volume data.

Images