| Autor: |
Surya, S., Ramamoorthy, S., Gopikrishnan, Arjun, Tanwar, Rudratej Singh, Rajeswari, D., Janani, K. |
| Předmět: |
|
| Zdroj: |
AIP Conference Proceedings; 2024, Vol. 3075 Issue 1, p1-8, 8p |
| Abstrakt: |
Gesture-based interaction is a popular method for human-computer interaction in the ubiquitous computing environment. With the advent of new and powerful IOT sensors and advancements in the field, there is a growing interest in leveraging these advancements to collect data from day-to-day activities. Several smartwatches come equipped with powerful accelerometer and gyrometer sensors, which are capable of measuring 3-dimensional data on a near continuous scale. Most smartwatches leverage machine learning classification models to categorize hand movements as triggers for the smartwatch application's behavior, such as accelerometer spikes and gyrometer shifts to turn on the fitness tracker in a smart watch. However, smartwatches battery life tends to be limited, and the gyrometer having to be kept on constantly does not help matters. Thus, this paper aims to address this issue by introducing a comparative analysis of hand gesture recognition using an accelerometer. Our study focuses on fourteen different hand gestures that involve motion of the subjects fingers as well as the wrist. The results obtained from the experiment confirm that the accelerometer signal can be used satisfactorily to track user motion and classify hand gestures, allowing the smart watch to conserve it's battery life. Initially, the analysis of signal to- noise ratio shows that the ACC signals are significantly more impacted by motion noise compared to the PPG signals. The findings of the research present a dependable basis for advancing the progress of gesture-based interactions in wearable technology. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
