Quality assurance of a gimbaled head swing verification using feature point tracking

Abstract

AbstractTo perform dynamic tumor tracking (DTT) for clinical applications safely and accurately, gimbaled head swing verification is important. We propose a quantitative gimbaled head swing verification method for daily quality assurance (QA), which uses feature point tracking and a web camera. The web camera was placed on a couch at the same position for every gimbaled head swing verification, and could move based on a determined input function (sinusoidal patterns; amplitude: ± 20 mm; cycle: 3 s) in the pan and tilt directions at isocenter plane. Two continuous images were then analyzed for each feature point using the pyramidal Lucas–Kanade (LK) method, which is an optical flow estimation algorithm. We used a tapped hole as a feature point of the gimbaled head. The period and amplitude were analyzed to acquire a quantitative gimbaled head swing value for daily QA. The mean ± SD of the period were 3.00 ± 0.03 (range: 3.00–3.07) s and 3.00 ± 0.02 (range: 3.00–3.07) s in the pan and tilt directions, respectively. The mean ± SD of the relative displacement were 19.7 ± 0.08 (range: 19.6–19.8) mm and 18.9 ± 0.2 (range: 18.4–19.5) mm in the pan and tilt directions, respectively. The gimbaled head swing was reliable for DTT. We propose a quantitative gimbaled head swing verification method for daily QA using the feature point tracking method and a web camera. Our method can quantitatively assess the gimbaled head swing for daily QA from baseline values, measured at the time of acceptance and commissioning.