Tracheal tree hierarchical extraction method combining multi-information fusion network and region growth

Abstract

The invention relates to a tracheal tree hierarchical extraction method combining a multi-information fusion network and regional growth. The method comprises the following steps: S1, acquiring a CT image of a lung, and preprocessing the CT image; S2, grading the preprocessed CT image set of the lung, and dividing the CT image set into two training sets, namely an overall tracheal tree and a tiny tracheal branch; S3, respectively sampling the overall tracheal tree training set and the tiny tracheal branch training set to obtain an overall tracheal tree training subset and a tiny tracheal training subset; S4, constructing a multi-information fusion segmentation model, and training according to the overall tracheal tree training subset; S5, constructing a voxel classification network model, and training according to the tiny trachea training subset; S6, sequentially inputting to-be-segmented image data into the trained multi-information fusion segmentation model and the trained voxel classification network model to obtain a preliminary tracheal tree; and S7, processing the preliminary tracheal tree by a geometric reconstruction method based on a center line to obtain a final tracheal tree. According to the invention, the classification accuracy is effectively improved.

Description

technical field [0001] The invention belongs to the field of medical image processing, and in particular relates to a tracheal tree hierarchical extraction method combined with multi-information fusion network and region growth. Background technique [0002] Due to the special anatomical structure and physiological function of the pulmonary trachea, lung diseases are closely related to its pathology. Therefore, segmenting the complete and accurate trachea from CT data plays an extremely important role in the preoperative diagnosis, intraoperative navigation and postoperative evaluation of lung diseases. The segmentation method of manual reading is affected by the number of slices and the complex tree structure of the trachea itself, which not only brings a huge workload to medical workers, but also easily leads to wrong segmentation. In the traditional segmentation method, manually extracting features relies on the knowledge guidance of professional scholars, and it is nece...

AI Technical Summary

Problems solved by technology

But at present, in terms of small trachea, deep learning methods still cannot get satisfactory results

Images