Features of Image Segmentation of the Upper Respiratory Tract for Planning of Rhinosurgical Surgery

摘要

Analysis of data of the SCT of the upper respiratory tract is the basis not only for the diagnosis, but also for the subsequent computer planning of rhinosurgical surgeries. For this, the process of automated segmentation of the anatomical structures of the nasal cavity is necessary, which significantly simplifies the preparation of geometric data for subsequent biophysical modeling. Segmentation of anatomical objects in this area should be carried out in two stages: selection of areas by density threshold values, subsequent marking of detected objects and post-processing -to eliminate artifacts and degenerate objects using morphological processing and logical filtering according to the geometric features of the analyzed objects. The arising segmentation errors are eliminated by the expert in an interactive mode.