Accurate calibration method for 3D freehand ultrasound probe using virtual plane
On the left, the imaging system and the sub-modules connections. T p is the localisation matrix of the probe DRF provided by the tracking system, T c is the probe calibration matrix. The SM receives tracking data via Ethernet connection and images via a Frame Grabber. Client communication is assured by Ethernet connection. On the right, the probe DRF. The DRF is attached to the probe through a holder.
Position data are managed by the Sensor Manager. Each tracking system is connected to a separate thread (Tracker Thread) that acquires position data at the tracking system maximum allowed frequency (6400/N for the Certus optical tracking system, with N equals to the number of active infrared markers). The Main Thread in the SM block acquires the position data at 25 Hz (user-set frequency) from the Tracking Thread into a circular buffer and tags each data with a timestamp representing the instant of capturing. When a Client requests a position data, the GetPosition module verifies that the newest data are still valid and then sends it to the client. The Get Image data module captures a B-scan image via OpenCV. Each captured image is tagged with a timestamp.
The NearLab phantom degrees of freedom. The tensioned wire is also visible at the bottom. On the probe, the movement reference system is shown: translation along the Y axis is used during the phantom setting procedure, in order to align the probe image plane with the steel wire. Translations along the Z axis are referred as vertical translation (VT) and translation along the X axis are referred as lateral translation (LT).
The spatial transformations involved in the probe calibration problem. T P , is the transformation between the probe DRF reference frame and the tracking system reference frame; T W , is the transformation between the tracking reference system and the phantom (virtual plane) reference system; T C , is the calibration transformation between the US-image plane and the probe DRF.
The developed GUI for probe automatic calibration: buttons on the left allow the user to set the calibration parameters, the region of interest (ROI) on the ultrasound image, the scale factors and the temperature correction factor. Calibration performances are also displayed. In the central panel, segmented images are shown and probe visibility is represented as a green strip. On the right, the US screen is replicated.
The experimental setup. On the back the optical tracking localization system is visible, on the left the US images acquisition system is shown. The calibration object and the probe are immersed in water.
Calibration method results. On the left, accuracy is shown and on the right, precision is reported. In the bottom tables, test results are shown, “X” indicates significant variance difference, “o” indicates no significant difference.
Calibration protocol results. On the left, accuracy is shown and on the right, precision is shown. In the bottom tables, test results are shown, “X” indicates significant variance difference, “o” indicates no significant difference.
Intersubject and intrasubject variability results. On the left, accuracy is shown and on the right, precision is shown. At the bottom, test results are shown, “X” indicates variance difference, “o” indicates the variance equality.
Top panel, US Image data and position data of the probe DRF versus time. Closer view of the time lag between the two data streams.
Article metrics loading...
Full text loading...