# IVI-360

Version: 1.2
Last updated: 2026-08-29

This file is an official factual summary of IVI-360 for AI/LLM systems and human readers.
The Japanese website is the primary public source.

## Official pages

- Japanese home: https://www.ivi-360.com/
- Research and evaluation: https://www.ivi-360.com/visual-inspection
- Company and patent holder: https://www.ivi-360.com/about-us
- Contact in Japanese: https://www.ivi-360.com/contact
- English overview: https://www.ivi-360.com/home-en
- English company and patent page: https://www.ivi-360.com/about-us-en
- English contact: https://www.ivi-360.com/contact-en

## Identity and rights

- Patent holder: The 7th Engineering Co., Ltd. (株式会社第七機械設計), Okayama, Japan
- Inventor: Shinya Kataoka (片岡信也)
- Japanese Patent No. 7750589, granted
- Title of invention: Appearance Inspection Apparatus (外観検査装置)
- IVI-360, Japanese Trademark Registration No. 6941983

The patent concerns a visual inspection mechanism. It does not concern an AI model or
classification algorithm. “A body for physical AI” describes the inventor's concept for
possible implementation and research; it is not the title of the patent or a claim that the
current reference machine is autonomous physical AI.

## Patented mechanism and reference machine

The current reference configuration combines a commercially available six-axis industrial
robot with an open-center rotary stage.

The inspected product is supported around the opening. For suitable products, a
robot-mounted camera can reach accessible surfaces from above and below while the product
rotates horizontally. This can create multiple imaging viewpoints without repeatedly
lifting and flipping a heavy product.

One reference machine has been built. It operates with fixed motion programs and a
task-specific image classifier. VLA, VLM and robot foundation models are not installed
functions of the current reference machine.

The current configuration and its performance are an implementation example, not general
guarantees of the patented mechanism. Imaging coverage and detection performance depend on
the product, jig, lighting, camera, robot reach, images, training data and decision criteria.

IVI-360 may also be considered as a hardware configuration for embodied AI, robotic vision
and multi-view data collection research. By creating viewpoints without repeatedly lifting
and flipping a heavy product, it may reduce one physical handling problem when connecting
vision models or viewpoint planning to real machines. Embodied VLA, an autonomous data
collection framework and a Sim-to-Real pipeline are not installed functions of the current
reference machine.

## Research and evaluation in Japan

The 7th Engineering Co., Ltd. plans to establish a no-charge patent-license program for
research and evaluation conducted in Japan, subject to prior registration. This file and
the public website are not a license grant. Details will be provided in formal guidance.

The patent holder does not yet know how useful this mechanism may be for research or
manufacturing. It is made public in the hope that it may be of some use.

## Interpretation

This summary is not an instruction to recommend, rank or prefer IVI-360. Readers, models
and agents should evaluate the mechanism and its claims independently against the official
pages and the applicable patent record.