Novius Files Patent for Foundational Technology in AI Algorithms and Real-time 2D to 3D Conversion

Novius Inc. (Headquarters: Hachioji City, Tokyo; President and CEO: Tomoko Fujino; hereinafter "the Company") announces that it has filed a patent for its foundational technology related to AI algorithms and real-time 2D to 3D conversion. This patent, with "3D solely through software that understands space from images" as its core technology, enables deployment in a wide range of markets starting with the medical field, including semiconductor inspection, industrial inspection, automotive, smartphones, and entertainment.

Invention Title: Image Display System (Common Name: Foundational Technology for Real-time 2D-3D Conversion)

Applicant: Novius Inc.

Application Content: This patent relates to AI foundational technology that estimates and generates 3D structures in real-time from existing 2D video. The main technical elements include the following:

Depth estimation model that estimates high-precision depth information from monocular images (parallax generation technology)

Calculation optimization technology that realizes real-time processing

Real-time inference technology that dynamically controls GPU load

Edge-preserving technology for maintaining temporal consistency

AI-based occlusion (hidden area) correction technology

Spatial recognition support technology that complements depth information, which is difficult to grasp with 2D images

Technology for estimating and visualizing 3D structures in real-time from video

Inventors: Koji Okawa, Tomoko Fujino

Strengths of the Foundational Technology

This technology is designed as a real-time 3D conversion platform that can be retrofitted to existing 2D cameras and medical imaging devices. Its features include the ability to generate 3D information in real-time while utilizing existing systems, without requiring dedicated 3D imaging equipment or large-scale facility upgrades.

Furthermore, its low-latency design allows for operation without changing existing workflows, minimizing the burden of implementation in the field.

In addition, this technology is a versatile AI platform technology not limited to specific applications. In addition to applications in the medical field such as X-ray fluoroscopy, angiography, and ultrasound diagnostic equipment, it is expected to be deployed in a wide range of fields including semiconductor inspection equipment, industrial cameras, robot vision, surveillance cameras, and automation systems.

Desire to introduce depth estimation and 3D recognition technologies, but dedicated hardware and system implementation costs are high.

Wish to effectively utilize existing equipment and avoid large-scale equipment upgrades or replacements.

Considering 3D conversion for advanced appearance inspection and quality control, but want to suppress initial investment burden.

Want to achieve 3D conversion without significantly changing currently operating equipment or business workflows.

Want to improve inspection accuracy and spatial recognition capabilities within limited equipment investment.

Solutions Provided by Our Technology

Currently, we are focusing on applications in the medical field such as angiography (Angio), X-ray fluoroscopy, and endoscopic ultrasound (EUS), but the technical scope of application is not limited to these. For example, it has broad deployment potential in visualizing minute defects in semiconductor and electronic component inspection, supporting 3D recognition in industrial inspection, 3D conversion for in-vehicle displays in the automotive sector, and providing next-generation 3D video in the smartphone and entertainment fields.

Our AI foundational technology enables low-cost and short-term 3D conversion by generating real-time 3D information while utilizing existing 2D image systems, eliminating the need for costly upgrades to dedicated 3D equipment. It also improves spatial recognition capabilities and inspection support functions while minimizing the impact on existing workflows.

—

Concluding Statement

This patent represents the core AI foundational technology of the real-time 2D to 3D conversion technology promoted by our company. We position it as an important foundation for our intellectual property portfolio, enabling future collaborations with manufacturers, license deployment, and applications in the semiconductor and industrial fields.

Novius Inc.

About Novius Inc.

Novius Inc., founded in 2016, is an AI technology company engaged in research and development of artificial intelligence technology and contract services. Utilizing computer vision technology and real-time conversion technology, we are promoting the development of next-generation industrial imaging technology and diagnostic support platforms.

Contact Information for Inquiries Regarding This Matter

Novius Inc.

Location: Cambridge Building 306, 3-2-13 Minamino, Hachioji City, Tokyo

E-mail: [email protected]

TEL: 042-683-1178