Motion Detection System with GPU Acceleration for Stereotactic Radiosurgery

概要

論文の詳細を見る
Stereotactic radiosurgery is a non-invasive method for the treatment of tumors that employs a narrow, high-energy X-ray beam. In this form of therapy, the target region is intensively irradiated with the narrow beam, and any unexpected patient motion may therefore lead to undesirable irradiation of neighboring normal tissues and organs. To overcome this problem, we propose a contactless motion detection system with three USB cameras for use in stereotactic radiosurgery of the head and neck. In our system, the three cameras monitor images of the patient's nose and ears, and patient motion is detected using a template-matching method. If patient motion is detected, the system alerts the radiologist to turn off the beam. We reduced the effects of variations in the lighting in the irradiation room by employing USB cameras sensitive to infrared light. To detect movement in the acquired images, we use a template-matching method that is realized with general-purpose computing-on-graphics processing units. In this paper, we present an outline of our proposed motion detection system based on monitoring of images of the patient acquired with infrared USB cameras and a template-matching method. The performance of the system was evaluated under the same conditions as those used in actual radiation therapy of the head and neck.