Enhancing Home Exercise Experiences with Video Motion-Tracking for Automatic Display Height Adjustment

Title of the Paper

Enhancing Home Exercise Experiences with Video Motion-Tracking for Automatic Display Height Adjustment

Paper Information

• Research Area: Human-Computer Interaction and Home Fitness System Design
• Keywords: Automatic Display Height Adjustment, Interactive Fitness Display, Head Tilt Angle, User Experience Design, NASA-TLX Evaluation

Research Background and Problem

• Identified Problems or Challenges:

  • With the growing demand for home fitness, how display devices can provide users with a comfortable viewing angle without interfering with fitness activities has become a significant challenge.
  • Poor screen design and usage can lead to musculoskeletal issues in the head, shoulders, and back.
  • Existing research primarily focuses on workplace scenarios, with a lack of studies specifically addressing optimized display design for home fitness.

• Significance:

  • In the context of home fitness, optimizing screen design to support users' musculoskeletal health is critical for improving exercise efficiency and user experience.

• Research Motivation and Related Work:

  • By integrating findings from human-computer interaction and automated furniture design, this study explores a technology capable of dynamically adjusting display height to promote proper posture and improve interaction experience.
  • The goal is to support comfortable head and neck angles while reducing unnecessary manual adjustments by users.

Solution

• Proposed Solution:

  • Developed an automatic screen height adjustment technology based on video motion-tracking data, dynamically modifying display height using skeletal information of the demonstrator.

• Innovative Aspects:

  • Utilized MediaPipe and Graph Neural Network (GNN) to perform real-time analysis of skeletal key points from the video demonstrator.
  • Optimized the YOLOv5 algorithm to accurately identify the boundaries of fitness equipment, providing a reference for shoulder height estimation.
  • Designed dynamic control curves to enable display devices to efficiently and safely respond to the demonstrator's movements.

• Implementation Steps and Techniques:

  1. Shoulder Height Estimation:
     • Extracted skeletal key points of the demonstrator using MediaPipe and predicted shoulder height with GNN.
     • Detected standard-sized fitness equipment (e.g., yoga mats) in the video and estimated shoulder height using the PnP algorithm.
  2. Control Curve Development:
     • Generated motion control curves for the display lift based on shoulder height, optimizing movement speed and work cycle.
  3. System Integration:
     • Integrated the height adjustment functionality into the video player, allowing the device to synchronize in real-time with the demonstrator's movements in the video.

Research Outcomes

• Specific Results:

  • The automatic height adjustment technology significantly reduced participants' head tilt angles, particularly upward tilt angles.
  • In user satisfaction studies, the automatic adjustment condition showed reduced mental, temporal, and effort demands, along with significantly lower frustration levels.
  • Users reported feeling more supported, with a notable improvement in interaction efficiency.

• Advantages over Existing Solutions:

  • The adaptive control curve enables smoother and more continuous display height adjustments, avoiding interruptions caused by manual adjustments.
  • The technology exhibits high real-time responsiveness, making it suitable for slower-paced fitness activities.

• Experimental or Evaluation Results:

  • In an experiment involving 30 participants, the head tilt angle under the automatic adjustment condition was significantly lower than under fixed height and manual adjustment conditions.
  • NASA-TLX evaluations indicated that participants' experiences with the automatic adjustment condition were superior to other conditions.

• Limitations and Future Directions:

  • The current technology is more suitable for slower-paced exercises and may not fully adapt to fast-paced or frequently changing shoulder height movements.
  • The study focused on a single type of cooldown exercise; future research should validate the technology's applicability to various fitness types.
  • The current adjustment mechanism does not account for individual user preferences; future work could incorporate user-specific body types and preferences for further optimization.
  • It is recommended to explore the impact of multi-directional (vertical and horizontal) display adjustments on user experience.

These findings provide significant theoretical support and technical guidance for designing interactive home fitness experiences, while also highlighting opportunities for improvement and expansion.