Market Context — Why This Technology, Why Now

The global demand for frictionless, secure access and personalized customer experiences is rapidly accelerating across all sectors. Simultaneously, privacy regulations are tightening, pushing for solutions that minimize personal data collection while maximizing utility. This technology offers a unique advantage by providing high-accuracy identification without explicit pre-registration, aligning with both operational efficiency goals and evolving privacy standards. It enables businesses to deploy advanced analytics and security measures with reduced user friction and compliance risk, driving adoption in smart cities, retail, and healthcare.

Key Competitive Advantages
01

Reduces operational costs by ~66% by eliminating the need for prior face registration, a requirement in conventional systems, significantly cutting initial deployment costs and operational effort. Annual operating costs could be reduced by approximately ~66%.

02

Achieves high-accuracy identification even for similar faces by utilizing a unique learning data generation technology that enables high-accuracy identification even among highly similar faces, substantially reducing misidentification risks and ensuring reliable person identification.

03

Generates learning data efficiently by automatically generating Triplet-format learning data, dramatically improving AI model training efficiency. This accelerates the post-deployment accuracy improvement cycle and enhances real-world adaptability.

Market Opportunity
🏢 Office & Facility Management
$350M–$1B globally (AI est.)
Demand is increasing for solutions that balance security and convenience, as systems requiring no prior registration significantly lower adoption barriers for access control and employee movement analysis.
Commercial real estate developers Large corporate campus operators Integrated security solution providers
🛍️ Retail & Commercial Facilities
$250M–$1B globally (AI est.)
There is growing need for high-accuracy person identification without burdening customers, for applications like customer behavior analysis and shoplifting prevention. This forms a foundation for providing personalized customer experiences.
Major retail chains Shopping mall operators Retail analytics software vendors
🏟️ Events & Entertainment
$200M–$1B globally (AI est.)
In environments with many temporary users, such as large-scale events, there is a demand for rapid and accurate person identification without prior registration for visitor management, lost child prevention, and VIP handling.
Event management companies Theme park operators Stadium and arena management groups
🏥 Healthcare & Elder Care
$200M–$1B globally (AI est.)
This technology could contribute to monitoring patients and residents, preventing wandering, and improving staff efficiency in hospitals and nursing homes. The absence of registration is a significant benefit, especially for the elderly or those requiring assistance.
Hospital system integrators Assisted living facility operators Medical device manufacturers with monitoring solutions
IP Defensibility — Why Competitors Can't Replicate This
What This Patent Covers

This patent broadly covers the technical scope with 10 claims, protecting the core algorithms for learning data generation and person identification without prior registration. Its successful prosecution, including overcoming examiner objections, indicates a robust and stable right, making it a strong strategic asset for licensees.

Competitive White Space

This patent primarily covers the AI learning data generation and identification algorithm. White space exists in developing specific hardware integrations, combining with other biometric modalities, or creating advanced behavioral analytics applications post-identification.

Economic Impact
~$50K/year estimated operational cost reduction per facility (est.)
estimated ROI · USD · AI analysis
ROI Calculation Logic

Implementing this technology eliminates the need for prior registration, a requirement in conventional facial recognition systems. For example, labor costs for annual new registrations and updates for 1,000 people (assuming 15 minutes per person at $20/hour (AI est.)) could be ~$5K/year (AI est.). Additionally, high-accuracy identification reduces misidentification response efforts (assuming 500 misidentifications per year, 10 minutes per case at $20/hour (AI est.)) for an estimated ~$16.5K/year (AI est.) in savings. Furthermore, automated learning data generation could reduce development man-hours by an estimated ~$11.5K/year (AI est.), totaling an expected annual cost reduction of approximately ~$33K (AI est.).

Speed to Market
7× faster than in-house development
The technology's learning data generation algorithm is well-established, with clear operational principles for each component (face detection, correspondence estimation, learning data recording, learning data set generation) detailed in the patent specification. This significantly shortens the conceptual design and basic research phases compared to in-house development, enabling rapid implementation into existing systems and prototype development. Early market entry is anticipated based on proven data.
Competitive Positioning

X: Ease of Deployment
Y: Identification Accuracy & Reliability

Business Models & Applications
💻 Software License Provision
This model provides the technology's learning data generation algorithm as software, integrating it into a licensee's existing person identification systems or camera infrastructure. It enables rapid deployment while minimizing initial investment.
☁️ SaaS-based Data Generation Service
This SaaS model processes image data provided by licensees in the cloud, generating and delivering high-accuracy Triplet-format learning data. It is suitable for companies with limited AI development resources.
🛠️ Integrated Solution Development
This model involves custom developing a person identification system, with this technology at its core, for specific industries (e.g., office access control, retail customer analytics) and offering it as an integrated solution combined with hardware.
Adjacent Application Opportunities
🏥 Healthcare & Elder Care
Contactless Monitoring System
In elder care facilities and hospitals, this could be adapted into a system that identifies residents or patients from camera feeds in rooms without prior registration. It has the potential to detect wandering or monitor individuals at risk of falls in real-time, reducing staff workload while enhancing safety for vulnerable populations.
🏙️ Smart City
Street Camera Movement Analysis
This technology could integrate with smart city street camera networks to analyze pedestrian flow during specific events or customer movement in commercial areas. While respecting privacy, it is expected to provide insights for urban planning and marketing strategies based on anonymized data, potentially improving traffic flow by 15-20% during peak hours.
🎭 Entertainment & Events
VIP/Specific Person Automatic Detection
This could be utilized as a system for automatically detecting VIPs or blacklisted individuals at concert venues, theme parks, or sports events without prior registration. It has the potential to balance smooth entry management with enhanced security, supporting safe operations without compromising the customer experience, potentially reducing security response times by ~25%.
Integration Roadmap — Estimated 12-Month Deployment
Phase 1: Proof of Concept (PoC) and Requirements Definition
Duration: 3 months
Conduct technical validation within the licensee's existing camera environment, evaluating the technology's learning data generation capabilities and identification performance. Concurrently, define specific deployment goals and detailed system requirements.
Phase 2: System Development and Prototype Construction
Duration: 6 months
Based on defined requirements, develop and integrate the technology's learning data generation module and person identification logic into the licensee's system. Build a prototype and conduct test operations in a small-scale environment.
Phase 3: Production Deployment and Operational Optimization
Duration: 3 months
Based on prototype validation results, proceed with full system production deployment and optimization. Establish a continuous improvement cycle for identification accuracy and data generation efficiency through on-site operations to maximize effectiveness.
Technical Feasibility
This technology is software-based, utilizing image data from existing camera systems for face detection, person ID assignment, and Triplet-format learning data generation, as specified in the patent claims. This allows for easy integration into existing surveillance or IoT camera infrastructure via software updates or API links, without requiring significant hardware upgrades, indicating a low technical barrier to adoption.
Success Scenario
Upon adoption, this technology could enable businesses, such as office buildings, to achieve seamless access control for visitors and employees without requiring prior face registration. This could automate reception tasks and enhance security, potentially reducing reception workload by approximately ~30% annually. In event venues, it is estimated to significantly boost customer satisfaction by virtually eliminating attendee registration wait times, contributing to higher repeat attendance rates.
Patent Record
APPLICATION NO.
特願2020-200432
REGISTRATION NO.
7589032
FILING DATE
2020/12/02
GRANT DATE
2024/11/15
EXPIRATION DATE
2040/12/02
PATENT HOLDER
日本放送協会
Examination History
2023年11月02日
出願審査請求書
2024年08月06日
拒絶理由通知書
2024年09月30日
意見書
2024年09月30日
手続補正書（自発・内容）
2024年10月15日
特許査定