Using a Somatosensory Controller to Assess Body Size for Size-Specific Dose Estimates in Computed Tomography

Area of Science:

• Medical Imaging
• Radiological Physics
• Biomedical Engineering

Background:

• Computed tomography (CT) is a vital diagnostic tool, but radiation dose management is critical.
• Size-specific dose estimate (SSDE) is used to determine CT radiation dose, requiring accurate patient body size measurements.
• Current methods for body size measurement can be time-consuming or involve radiation exposure.

Purpose of the Study:

• To evaluate the efficacy of a somatosensory controller (Microsoft Kinect) integrated with a CT scanner for real-time measurement of patient body dimensions.
• To compare the accuracy of Kinect-based measurements with conventional methods like F-pucks and CT imaging.
• To assess the potential of Kinect for improving the accuracy and efficiency of size-specific dose estimate (SSDE) calculations.

Main Methods:

• A Microsoft Kinect sensor was utilized to capture depth images of 108 individuals.
• A grayscale-to-surface height conversion curve was developed for anterior-posterior diameter (APD) estimation.
• Measurements of APD, lateral diameter (LATD), and effective diameter (ED) using Kinect were compared against F-puck and CT image-derived values.

Main Results:

• Kinect measurements showed low mean absolute differences compared to F-puck (APD: 5.2%, LATD: 1.3%, ED: 2.5%) and CT images (APD: 8.8%, LATD: 2.6%, ED: 5.0%).
• The system demonstrated high accuracy in estimating patient body dimensions.
• The Kinect system offers a simple, low-cost, radiation-free, and automated solution for body size assessment.

Conclusions:

• The Kinect-based somatosensory controller is a viable and accurate alternative to traditional methods for measuring patient body size in CT examinations.
• This technology can enhance the precision of size-specific dose estimate (SSDE) calculations, contributing to safer radiation practices.
• The integration of Kinect offers significant advantages for real-time, non-invasive patient sizing in radiological settings.