FIGURE 7.

Deep learning-based automatic targeting application for Patient 1. (A) Input image for targeting. (B) Semantic segmentation results and automatic targeting results. Bilateral ends of the bent edges of the white line are automatic targets calculated using the method shown in Figure 3. (C) Semantic segmentation imaging and automatic targets superimposed on the input MRI. (D) Automatic targets superimposed on the input MRI. (E) Orange dot indicates the deep learning-based automatic target in the left posterior subthalamic area, while the green dot is the indirect target. The cyan dot is the post-operative location of the electrode. (F) Planned trajectory and target (red dot and line) and location of the post-operative electrode (cyan dot) superimposed on the preoperative MRI. (G) Planned trajectory and target (red dot and line) and location of the post-operative electrode (cyan dot) superimposed on the post-operative CT image.