Tech Blog

Validation of REMO's AI-Based Posture and Gait Analysis Technology

REMO has established a reliable technology foundation by validating its AI-based posture and gait analysis technology using standard camera images through performance metrics, research, clinical applications, and real-world deployment cases.

3D Pose Estimation

REMO has developed AI technology that analyzes human posture, gait, and movement using standard camera images, and has validated this technology through performance metrics, research, clinical applications, and real-world deployment cases. Performance data, academic papers, hospital application cases, patents, and government-funded project records show that REMO's technology has evolved through actual validation and execution, rather than remaining at the level of a simple concept.

REMO's AI posture analysis technology is designed to assess nine categories of postural imbalance based on front and side-view images. In addition, REMO has advanced technology that predicts key body measurements such as waist circumference, hip circumference, thigh circumference, arm length, and leg length using only standard camera images. Its gait analysis technology has also evolved to quantify data points such as gait speed, cadence, step width, step length, stance phase, swing phase, left-right balance, and lower-extremity joint angle changes from standard video footage.
REMO's technology has been validated through specific performance indicators. Posture analysis showed 98 percent angle measurement accuracy compared to commercial equipment, body circumference analysis achieved an average prediction error within 3 cm, and gait analysis recorded a joint measurement error of 32.73 mm MPJPE (World's highest level 30 mm). In addition, REMO has further strengthened the credibility of its gait analysis technology by obtaining medical device manufacturing approval in Korea and validating its performance against commercial equipment.

REMO also utilizes not only real-world data but also its own motion musculoskeletal simulator and synthetic biomechanical data pipeline. Based on a training framework that combines real data and motion synthetic data, REMO continues to improve both the precision and scalability of its analysis technology.
This technical validation also extends into research, academia, and real-world application. In a paper validating a deep learning-based simplified posture analysis technology, accuracy and reliability were evaluated against 3D motion analysis systems. In hospital application cases, monitoring of gait changes before and after treatment and improvements in joint angles were presented.
REMO's technology has progressed beyond the research stage into a broad range of outcomes and execution records. Achievements in AI competitions hosted by NIPA, a Korean government-affiliated agency, 14 patents, venture business certification, corporate research institute certification, registration as an official NIPA technology supplier for six AI technologies, and execution of 16 Korean government-funded projects demonstrate that REMO has continuously validated and advanced its AI analysis technology.
REMO continues to enhance its standard camera-based posture, gait, and motion analysis technology so it can be applied more broadly across real-world environments. Going forward, REMO will continue to advance trustworthy AI analysis technology based on research, validation, and practical field experience.

If you would like to learn more about REMO's technology validation and real-world application cases, please feel free to contact us. We would be happy to discuss implementation models and potential areas of collaboration.