The collaborative effort, involving Ninenovo, Professor Cheng Chen’s team, and Mianyang Central Hospital, examined 1,810 participants to determine if bilateral signals yield better health data. Conventional wearable systems typically assume that readings from the left and right sides of the body are interchangeable. This research refutes that assumption, revealing that signal differences between hands serve as a key indicator of measurement reliability.
By synchronizing dual-hand tracking, researchers successfully reduced the mean absolute error for heart rate from 10.65 bpm to 3.21 bpm. When signal differences remained below specific thresholds, accuracy improved further to 2.59 bpm. These gains remained consistent across 12 different AI model architectures. While improvements in blood pressure estimation were more modest—dropping systolic error from 13.81 to 12.36 mmHg—the results underscore that sensor placement and signal consistency are as vital as raw data volume for the future of digital health diagnostics.

Comments (0)
No comments yet. Be the first!