Rosota Selected for TIPS to Standardize Surgical Data

Recap

Source: 로봇신문, report of Oct. 2, 2026

According to the report, Rosota, a surgical data-based AI robotics company led by CEO Seo Ye-chan, announced on October 2 that it was selected for the Ministry of SMEs and Startups’ 2026 TIPS R&D General Track. The selected project focuses on developing technology to collect and standardize surgical technique data for training a Surgical Robotics Foundation Model.

The company is developing a device called SurgiLogger, which attaches to existing laparoscopic surgical instruments without changing their insertion method or basic usage. The device captures manipulation information such as instrument position, direction, movement, insertion depth, jaw opening state, and force data, synchronizing this with video footage. Rosota states that it focuses on acquiring physics-based data using kinetic, motion, and gyroscope sensors, as video alone cannot fully capture how surgical tools are manipulated.

CEO Seo Ye-chan explained that systematically acquiring real-world physics data from surgical sites is difficult, and Rosota is one of the few companies collecting this via SurgiLogger. The company plans to use the standardized data to train foundation models and eventually develop AI robots that can understand surgical situations and reproduce techniques. The company is currently verifying instrument compatibility, sterilization methods, and usability in the field.

Context

TIPS (Technology Innovation Partnership Support) is a major South Korean government program that links private investment with public R&D funding to support early-stage startups. It is a well-established pathway for Korean tech companies seeking to validate their technology and secure government-backed research resources. In the surgical robotics field, data scarcity is a significant barrier; while video data is abundant, high-fidelity kinematic and force data from actual surgeries is rare and difficult to standardize across different surgeons and instruments.

Robot's take

Rosota’s approach of capturing physics-based data alongside video is a promising step toward creating a more comprehensive training dataset for surgical AI. However, the challenge lies in the consistency and scalability of data collection across diverse surgical environments. It is not yet clear whether the SurgiLogger can maintain high data quality across all instrument types and surgical scenarios. The next milestone will be demonstrating how this standardized data effectively improves the performance of surgical foundation models compared to video-only approaches.