11.3 AI Motion Analysis Technology

LILLIUS AI motion analysis technology undergoes three progressive phases:

1. In Phase 1, we focus on data collection and the enhancement of one-person motion analysis technology.

2. In Phase 2, we introduce personal sports challenge services, offering exercise and friend recommendations, subscription options, gift-giving features, and metaverse fan services.

3. Finally, in Phase 3, we implement a group-type challenge service that allows real-time linkage with exercise evaluation scores.

<AI motion analysis technology phases>

#1. Advancement of one-person motion analysis technology ⇒

#2. Personal sports challenge service (exercise and friend recommendations, subscription options, gift-giving, metaverse fan service, etc.) ⇒

#3. Group-type challenge service (real-time linkage with exercise evaluation scores)

11.3.1 Refinement of Technician Skills is the core feature

• The features are enhanced by stabilizing bone extraction information and utilizing core joint selection technology.

• Additionally, hybrid features are integrated, combining the shape of core postures with the dynamics of core motions.

11.3.2. Motion Accuracy Analysis Technology

• Modified Dynamic Time Warping is employed to analyze motion accuracy in each detailed section

• Choreography motion differences for each joint are analyzed to provide a detailed motion error analysis.

11.3.3. Accuracy Analysis Performance Improvement and Utilization Technology

• The similarity between the user and the correct action is calculated using correlation analysis between the features in each section. This process allows for diversification of evaluation criteria through parameter adjustment.

= Evaluation section length, weight by body part, pose/dynamics/timing score weight, etc.

LILLIUS aims to expand its business using user exercise data through the AI motion analysis system, focusing on the following areas:

1. Precise motion analysis to measure the user's follow-up accuracy.

2. Expanding the data business based on diverse exercise motion data.

3. Establishing the foundation for a personalized exercise recommendation service.