GAZIRU.z (Individual Identification)
A server product for tagless individual identification that pinpoints a single item by its fine surface patterns.
Try it in your browser
Tell two identical-looking images apart, on the spot
Upload two images and see a similarity score showing whether they are the same individual item. Sample images are included, so you can try it right away without photos of your own. Access keys are free and issued through our contact form.
Revolutionary technology that identifies individual items using only surface images
GAZIRU.z (Individual Identification) is a revolutionary service that recognizes unique surface patterns invisible to the human eye and identifies individual items.
No need for QR codes or RFID tags. Simply photograph objects with a camera to instantly match unique individual information.
Since no modification to products is required, it can be used in various fields including traceability management of materials and parts, and genuine product verification of branded products.
Main features
- Register and match individuals (one at a time or in batch)
- Narrow matching by specified group
- Alignment feature (corrects position and orientation differences on the server side; enabled by default)※2
- Operating modes: standard / register-only / match-only (switchable per data group)
- Separating match targets: partition the matching data by unit — such as lot or manufacturing month — and load only what's needed at match time
- Distributed configuration: shard distribution (scales the pool) and replicas (redundancy)
- RESTful API (V5 API and V4-compatible API), command line, and a web browser admin console
- Separated management of multiple business units and data groups, OpenID Connect authentication, per-data-group backups, and operation logs
Deployment options
Server product (on-premises)
Installed on a Linux OS (RHEL family), so it can be deployed on the factory floor, in-house, or on a closed network. A minimum configuration is roughly 4 cores, 8GB memory, and 10GB disk (excluding data).
Cloud service
Offered as monthly plans tiered by the number of registered items. From evaluation through production systems, you can get started with low upfront cost.
Client product
A client for operating the API-based server product from a PC at hand. It also includes image capture, so you can easily begin verifying whether individual identification is feasible. In production systems, the client side is typically built to fit your use case, such as integrating with the imaging setup or triggering the next action based on match results.
License
On-premises licensing is based on the number of physical cores, with a choice of perpetual (one-time purchase) or term licenses. Adding cores increases matching parallelism for faster response. Choose according to your requirements and scale.
GPU option
An option that uses a GPU to accelerate processing, chiefly the alignment step, whose processing time depends on the image content.
The imaging setup (camera, lighting, fixtures), and for on-premises deployments the hardware such as servers and network equipment, is assumed to be prepared by the customer or a system integrator. Depending on conditions, we can also build the system for you. For guidance on shooting conditions, see the article below.
How it works
Mechanism: Identifying Objects with Image Recognition
The mechanism of GAZIRU.z is very simple.
Registration
Photograph the surface of items to be identified and register them as computable numerical data in the database.
Matching
When photographing items to be identified with a camera, data is read.
Determination
The read data is matched against information registered in the database to instantly identify which individual it is.
This technology is realized by analyzing minute patterns and textures on object surfaces with proprietary algorithms. It is robust against focus drift and lighting variations, enabling highly accurate and stable identification.
Identifying Individuals by Unique Surface Patterns
GAZIRU's individual identification technology captures microscopic patterns and textures that exist naturally on product surfaces at the micrometer (μm) level, or unique patterns that occur incidentally during the manufacturing process. Simply by taking a photograph of the surface, AI extracts features from the captured content and matches them against a database to identify individual objects with extremely high precision.
Unique Technology
Our proprietary image processing technology enables tag-free, fast and accurate individual identification across a wide range of targets simply by taking photographs
Non-invasive & Tag-free
No processing or tagging of products required. Preserves aesthetics and enables individual management even for ultra-small parts that were previously difficult to manage
High Precision & Reliability
Recognizes minute differences and keeps misidentification low
High-speed Processing
Even on general-purpose servers, matches against a pool of one million items in around a second.※1
Versatility
Works with any target without requiring object-specific learning or tuning. Supports diverse materials including metal, plastic, paper, fabric, from flat to three-dimensional objects
API Integration
Easy integration with customer systems through engine interfaces (RESTful WebAPI) provided by cloud services or on-premise products installed on servers
Matching is judged by a score. Scores are distributed so that different individuals score low and the same individual scores high, and the judgment is made against a threshold. For more, see the article below.
Performance at production scale
How many individuals can it search, and how quickly? These are results of in-house tests at production scale, including registration and continuous operation.
Matching response
Matches against a pool of roughly 1 million items in around a second. Matching is processed in parallel across distributed workers, and response time shortens almost proportionally as parallelism increases. Partitioning the matching pool by unit, such as lot or manufacturing month, is supported as standard, and the design scales to a 10-million-item range.※1
Registration throughput
Completed a bulk registration of 1 million items in about 9 hours (roughly 110,000 items per hour, enough throughput to register a full day's production overnight in one batch).※3
Continuous operation
In a continuous-operation test equivalent to one year of parts traceability, processed about 2.5 million requests without errors.※4
Implementation Benefits: The Future of Business with GAZIRU
Quality Management
Track manufacturing history and assembly processes of each part to achieve strict traceability.
Genuine Product Verification
Prevent circulation of counterfeit and imitation products by recording individual information of branded and high-value products.
Supply Chain Management
Prevent product substitution and loss during distribution stages to achieve highly transparent logistics.
Operational Efficiency
Eliminates the need for label and tag attachment work, leading to cost reduction and improved work efficiency.
Application Scenarios: Where GAZIRU.z Excels
Manufacturing Industry
Manufacturing history management and recall response for small parts and complex-shaped components
Logistics & Distribution
Individual management of branded and high-value products, verification of identical items during returns and exchanges
Service Industry
Inventory management of rental and lease items, recording of inspection and maintenance history
Case studies
Apre Co., Ltd. — In the secondary market for luxury brand goods, GAZIRU.z links appraised individual items to the physical goods on hand, streamlining condition-assessment work.
Dadway Co., Ltd. — Adopted GAZIRU.z as a countermeasure against counterfeit products for its baby carriers (Ergobaby, Japan-exclusive new products).
For Those Considering GAZIRU.z Implementation
Please feel free to contact us for detailed specifications, implementation images, case studies, estimates, and more.
A browser-based matching demo is also available. Access keys are free and issued through our contact form.
Matching demo environment (compare two images and see the score)
Notes
※1 Time to find one individual among 1,000,000 registered items: about 1.5 seconds with 3-way parallelism (without alignment; three machines with 8 physical cores / 64GB each within a public cloud platform; measured in-house in July 2026 on a GAZIRU.z V5.0.0 development build). Results for 4-way parallelism and above, or more than 1,000,000 items, are estimates based on the scaling trend confirmed by measurement. Actual response time varies with the equipment, network configuration and other conditions. With alignment enabled, response time also depends on the image content. A GPU option is available to speed up matching when alignment is enabled.
※2 The alignment feature automatically corrects, where possible, differences in shooting position and orientation at match time, and assists identification accuracy by raising the score only for the same (correct) individual.
※3 Batch registration test of 1 million items (no image storage; measured in-house in June 2026 on a GAZIRU.z V5.0.0 development build).
※4 Continuous-operation test assuming traceability for a line producing 5,000 units per day: one year (240 business days) time-compressed into roughly 4.8 days of continuous operation. Approximately 1.22 million individual items and 2.5 million requests (tested in-house in August 2026 on a GAZIRU.z V5.0.0 development build).