Syncing the IMU data with its respective US imaging

[nickawwas]

I am collecting tracked US data using the Clarius L15 and L20 probes. I have a few questions for syncing the timestamps between the US images and IMU data.

1. What is the latency between capturing the frame and processing the B-mode image? And does framerate affect this latency?

2. Does the timestamp associated with each frame provide the time of capture or time after processing?

3. Are the timestamps provided from the IMU the time in nanoseconds and how do I associate them with their respective US images?

4. What is the current reconstruction error from using the IMU for short sweeps?

[clariusk]

1. everything is time stamped as soon as the DMA completes (ie. full frame acquired and sent to memory), for imu: when the interrupt completes and fifo is flushed for N imu samples. so limited latency in terms of true acquisition vs processing
2. as per above
3. same as the images, nanoseconds
4. n/a we have not measured

[nickawwas]

Thank you for your answers.

I'm obtaining different timestamps for the videos and the imu data. The video file is called capture_1_2026-07-31T15-23-42.dcm while its respective capture is called 2026-07-31T19-23-41+0000_imu.yml. The first timestamp for the US video capture is 152342032000 and the second timestamp for the IMU data is 195212841780; these do not appear to match. Could you explain these discrepancies?

[clariusk]

yes, DCM does not line up as these are captured in the app, which is not in sync with the probe, to align images and imu, use the raw data store that is buffered and sent back to the app.