Wedge Calibration

Stradx offers this spatial calibration option to those users who have a mechanical alignment phantom, as described in the following technical report:

A.H. Gee, N.E. Houghton, G.M. Treece and R.W. Prager. 3D ultrasound probe calibration without a position sensor. Technical report CUED/F-INFENG/TR 488, Cambridge University Department of Engineering, August 2004.

Download the technical report now (14.0 MBytes PDF, 21 pages).

To calibrate this way, first open the `Preview' window and click the `Wedges' toggle button. The `Wedge calibration' panel will pop up. Mount three pairs of wedges on the calibration frame: one at the centre of the top wire, the other two near the bottom corners of the cropped B-scan, as it appears in the `Preview' window. Put the frame in the water tank, fill the tank with water at around 50 degrees Celsius, and mount the probe on the phantom.

The principle of operation is that the ultrasound beam is manually aligned with the plane of the wires, using the wedges to guage the local alignment. First, adjust the x and y stages until the top wedge-pair is near the top centre of the B-scan in the `Preview' window. Set the ultrasound focus for the top wedge-pair. Set the time-gain control to highlight the wedges and suppress the rest of the B-scan. Adjust the z stage until the echos from the two top wedges appear at the same depth. The ultrasound beam is now roughly aligned with the top wedge-pair.

Next, adjust the yaw and pitch stages until the bottom wedge-pairs are aligned with the beam. The echos from the left and right wedge in each pair should look more or less the same, without any obvious asymmetry between left and right. The ultrasound beam is now roughly aligned with all three wedge-pairs: to fine tune the alignment, we're going to use some image processing tools.

Use the left mouse button to click and drag a bounding box around each wedge-pair. Click first at the top centre of the wedge-pair, then drag towards one of the bottom corners. Two indicators of wedge alignment appear. The first comprises a pair of cyan horizontal lines, each centred on a wedge's echo. As you adjust the z stage, you should find that the echos, and hence the cyan lines, move up and down with respect to each other. The ultrasound beam is centred on the wedges when the lines are level.

The second alignment indicator comprises a number, displayed in green near the bottom of the bounding box, quantifying the difference between the reflections in the left and right halves of the bounding box. Small numbers indicate small differences and therefore good alignment. Towards the top of the bounding box is another green number, which is the smallest difference found so far.

The Preview Window

The horizontal cyan lines work well when the ultrasound beam is narrow compared with the size of the wedges. This is usually the case for the top wedge-pair, less so for the bottom pairs. When the beam is so wide that the wedges lie entirely inside it, like at the bottom of the B-scan in the picture above, you'll find that the cyan lines are not helpful and you should instead use the green numbers: you're looking to make them as small as possible.

Start by fine tuning the alignment of the top wedge-pair, by adjusting the z stage until the cyan lines are level. To calibrate the probe, the system needs to know the precise location of this pair's central axis. To find this, move the z stage away from its aligned position until the bounding box's red centre-line becomes dashed. Since the wedges now show up as two distinct blobs, the system can analyse the echos to find their central axis. After a few seconds, the centre-line turns green. Now adjust the z stage back towards its aligned position and out the other side, until the centre-line is dashed again. After a few seconds, the centre-line turns cyan. The central axis is now fixed. Return the z stage to its aligned position.

Next, set the ultrasound machine's focus for the bottom wedges and fine tune their alignment. Unless you have a "2.5D" probe with a very narrow beam, you'll probably want to use the green difference indicators and not the cyan lines. Adjust only the yaw and pitch stages until both green numbers are as small as possible (or both sets of cyan lines are level). Do not touch the z stage, since this will disturb the alignment of the top wedge-pair. When you're finished, the `Preview' window should look something like this.

The Preview Window

Finally, we need to segment the wires supporting the wedges. Each pair of wedges is actually supported by two wires: it's the upper of these two that we're interested in. With the middle mouse button, click on the upper wire to the left and right of the top wedge-pair. The system draws a red line through the two points. Next, again with the middle mouse button, click on one point on the upper of the two wires supporting the bottom wedges: the system draws a red line through this point, parallel to the top line. If you click in the wrong positions, simply click again with the middle mouse button. Likewise, if you get the wedge bounding boxes wrong, you can delete a wedge with the right mouse button, then re-define it. The `Preview' window should now look something like this.

The Preview Window

One last detail is to tell the system whether the B-scan is flipped left/right or not. At the top of the `Preview' window, in red, are the letters `L' and `R'. Dip your finger into the water tank and wiggle it about somewhere above the bottom left wedge-pair (left as your face the phantom). Does your finger appear at the left of the B-scan in the `Preview' window? If yes, then the B-scan is not flipped and there's no need to do anything. But if your finger shows up at the right of the B-scan, then move the mouse into the `Preview' window and press the space bar. The `L' and `R' will swap around, as in the picture above, to indicate a flipped left/right B-scan.

The Wedge Panel

Now, read the micrometers and enter the values into the `Wedge calibration' panel using the five sliders. You also need to tell the system which wires the wedges are mounted on. To do this, simply count down the holes from the top of the frame. Remember that each wedge is supported by two wires, and it's the upper of the two we're interested in. The top wedge-pair is always mounted on the top two wires: the lower wedges are mounted to suit the depth range of the probe you're calibrating. In the example above, the bottom wedges were mounted on wires strung between the 7th and 8th holes from the top of the frame: hence, the slider is set to 7.

With all the sliders set, you can now press the `Calculate calibration' button to work out the spatial calibration. This button is insensitive until the panel has all the information it needs from the `Preview' window: the two wires and the location of the top wedge-pair. The calculated calibration parameters are immediately displayed in the calibration controls window. Finally, save your results in a calibration file.