An improved A* algorithm based on mean shift in bronchial environment

Lung cancer is one of the most prevalent malignancies with the highest mortality rate worldwide, posing a significant threat to human health. With advancements in CT screening technology, the detection rate of pulmonary nodules has increased, making bronchoscopy an essential tool for the early diagnosis and treatment of lung cancer. Bronchial path planning plays a crucial role in bronchoscopic interventional surgery and has become a key research focus in this field. To address the prolonged computational time in bronchial path planning, an improved A* algorithm based on the mean shift method is proposed. First, a node optimization strategy is introduced by incorporating the mean shift algorithm. Leveraging its mean shift properties, this approach ensures that nodes move toward the bronchial center and away from the airway wall, thereby enhancing path safety. Second, a variable step-size sparse spherical node expansion method is designed to expand the search space and reduce computational time. Additionally, a dynamic step-size adjustment mechanism is employed to improve adaptability by accommodating the geometric variations of different bronchial regions, thereby enhancing the efficiency of path planning. Furthermore, a cubic B-spline curve is applied to smooth the planned path, ensuring its feasibility. Finally, the effectiveness and superiority of the proposed improved A* algorithm based on mean shift, the traditional A* algorithm, and the Voronoi centerline method are validated using lung CT data.