Semi-automatic ct image segmentation using random forests learned from partial annotations

Human tissue segmentation is a critical step not only in the process of their visualization and diagnostics but also for pre-operative planning and custom implants engineering. Manual segmentation of three-dimensional data obtained through CT scanning is very time demanding task for clinical experts and therefore the automation of this process is required. Results of fully automatic approaches often lack the required precision in cases of non-standard treatment, which is often the case when computer planning is important, and thus semi-automatic approaches demanding a certain level of expert interaction are being designed. This work presents a semi-automatic method of 3D segmentation applicable to arbitrary tissue that takes several manually annotated slices as an input. These slices are used for training a random forest classifiers to predict the annotation for the remaining part of the CT scan and final segmentation is obtained using the graph-cut method. Precision of the proposed method is evaluated  on CT datasets of hard tissue including tibia, humerus and radius bones, mandible and single teeth using the Dice coefficient of overlap compared to  fully expert-annotated segmentations of these tissues.