Breaking Down Couro AI’s Run Score

Couro AI is the latest in run analysis technology. We’ve developed a run score that summarizes running performance for a given individual based on run symmetry.

Running Symmetry Score (RSS)

Based on joint movements over time, we are measuring joint symmetry during non-transitory phases for both the left and right sides of the body. A primary assumption is that imbalances in joint movements are indicators of imbalance that can further lead to injury if unchecked/uncorrected.

To evaluate running symmetry between the left and right sides, we calculate the Pearson Correlation Coefficient (PCC) of the two waveforms for the following rationale:

• A value of -1 indicates that there’s perfect symmetry between the two waveforms, while any other value will indicate there is some deviation.

• Thus, if we take the absolute value of the PCC of these two waveforms, we get a scaled value that tells us how different the left and right sides are.

  • Note: this interpretation does fall apart if there is a correlation between the two wave forms (the person is swinging their arms and legs perfectly together, like in a jump).

The RSS will be:

RSS = abs(PCC(\phi_{left}, \phi_{right}))

Where 1 indicates perfect symmetry between the left and right side swings, and 0 indicates random joint movement. The model will also calculate the RSS over 4 joints and two sides:

• shoulder

• hip

• elbow

• knee

The reports will include each individual symmetry score and a composite symmetry score that will be used on the final athlete score.