Contract development of AI systems that run on smartphones
We develop practical AI systems for smartphones, centered on OCR, image recognition, and on-device AI. We support high-speed, high-accuracy implementations for offline environments and field operations.
We develop practical AI systems for smartphones, centered on OCR, image recognition, and on-device AI. We support high-speed, high-accuracy implementations for offline environments and field operations.
High-speed OCR and barcode recognition
Offline AI design for field operations
AI integration into Android / iOS apps
Support from PoC to production improvement
High-speed OCR and barcode recognition
Offline AI design for field operations
AI integration into Android / iOS apps
Support from PoC to production improvement
ISSUES
Do you have any of these challenges?
You want to improve operations with AI but the barrier to adoption feels high
Cloud AI is causing delays depending on network conditions
You want the system to work offline in the field
You want to embed AI functions into a mobile app but do not know how
You completed a PoC but cannot reach production quality
These issues can be addressed with AI that runs directly on smartphones.
These issues can be addressed with AI that runs directly on smartphones.
SERVICES
We develop AI systems that run on smartphones
Nefrock provides contract development of AI systems that run on smartphones. We can support the entire process from requirement definition and PoC to implementation and improvement.
Nefrock provides contract development of AI systems that run on smartphones. We can support the entire process from requirement definition and PoC to implementation and improvement.
OCR system development
OCR result
Date
2023-01-01
Control No.
A-10293
Barcode and label reading
Slip reading result
Part No.
TA02G
Slip No.
1234-5678-9012
Image recognition engine development
Image recognition for your needs
Auto detection
Detect target objects automatically
Auto classification
Classify results to your needs
Custom model
Build proprietary detection and classification models
Adding AI features to existing apps
Operational efficiency with AI
Auto input
Input assistance
Auto classification
Classify data automatically
Custom model
Customization for your use case
Why run AI on smartphones?
Running AI on smartphones enables operations that are difficult to achieve with cloud AI alone.
Typical cloud AI
Requires a connection and depends on the environment
Depends on network conditions
Cloud transmission is required
Response tends to be slow
Cannot be used offline
On-device AI
Runs entirely on the phone. Works in the field.
Processes on the spot without depending on the network
Easy to operate without requiring cloud transmission
Fast response suited to field work
Stable operation even offline
On-device AI is effective for field operations and real-time use cases.
STRENGTHS
Why Nefrock
Our strength is not only in AI model development, but in making systems work in a form that is actually usable.
01
Design focused on high-speed processing on smartphonesDesign focused on high-speed processing on smartphones
02
Strong expertise in OCR and recognition algorithmsStrong expertise in OCR and recognition algorithms
03
End-to-end support from AI development to app implementationEnd-to-end support from AI development to app implementation
04
Flexible customization with our proprietary AI engineFlexible customization with our proprietary AI engine
05
Support from PoC to production improvementSupport from PoC to production improvement
CASE STUDIES
Case studies
Here are examples of how smartphone AI and recognition technologies have been integrated into real products and workflows.
Product
EdgeOCR
A high-speed OCR engine that runs on smartphones
An OCR engine optimized for logistics and field operations. It performs text and barcode recognition directly on smartphones and supports stable operation even in offline environments.
Case study
Virtual try-on for JINS
Customer experience powered by image generation and conversion
A virtual try-on example that lets users try on different glasses while keeping their own glasses on. Real-time image conversion creates a new purchase experience in stores and online.
Case study
Glasses fit scoring for JINS
Product recommendations powered by a recommendation engine
A case where the evaluation criteria of store staff were learned to provide fit scoring and ranked recommendations linked with virtual try-on. It supports product suggestions while preserving the quality of customer service.
Case study
Age estimation AI for Toshiba Tec
Face analysis and ID OCR that complement human judgment
A case where we developed, end to end, an edge AI that estimates a person's age from camera footage — from model training to the on-device SDK. Combined with ID OCR, it supports age verification that complements human judgment, running fast on the device.
Case study
Verification system for Takebishi
Preventing mix-ups on site with OCR and master matching
A case of a verification app that supports mix-up prevention on manufacturing and logistics floors. It uses OCR to read printed codes and text and automatically matches them against registered master data, instantly preventing operational mistakes.
Case study
Inbound/outbound inspection system for Sanritsu
Systematizing complex rules and know-how to achieve zero mis-shipments and freedom from reliance on individuals
A case where, at the warehouse of Sanritsu, a comprehensive logistics services provider, an inspection system with the functions logistics sites had been looking for resolved challenges around inconsistent work quality and knowledge transfer that depended on skilled staff.
Case study
Verification app for SBS Toshiba Logistics
Verification that handles diverse labels
A case of introducing an OCR app to assist visual checks and manual data entry during receiving and shipping. It handles diverse label formats, preventing errors and improving work efficiency.
Case study
Import/export verification app for MD Logis
Versatile handling of complex matching rules
A case of an app that supports the verification of work orders, cargo, and labels in import/export operations. With versatile handling of complex matching rules, it reduces human error and prevents verification omissions regardless of an operator's experience.
Use cases
Smartphone AI is used in a wide range of field operations.
Label and printed text reading in manufacturing sites
Barcode and text recognition in logistics sites
Inventory and stocktaking efficiency improvements
Product information reading in retail sites
Image recognition for inspection and checking tasks
Adding AI features to existing business apps
PROCESS
From consultation to development
You can consult with us even before the development scope is fixed.
You can consult with us even before the development scope is fixed.
STEP1
Hearings on issues and requirements
STEP2
Organizing feasible approaches and proposal
STEP3
PoC development
STEP4
App implementation and AI integration
STEP5
Evaluation, improvement, and operational support
We improve accuracy and speed step by step until the solution is ready for real operations.
For those who want to know more about OCR
For those who want to know more about OCR
If you are mainly considering text recognition and barcode recognition, we can also introduce EdgeOCR, an OCR engine that runs at high speed on smartphones.
If you are mainly considering text recognition and barcode recognition, we can also introduce EdgeOCR, an OCR engine that runs at high speed on smartphones.
High-speed processing on smartphones
Offline operation
Optimized for field operations
High-speed scan in 0.1 sec
No-code configuration
High-speed processing on smartphones
Offline operation
Optimized for field operations
Talk to us about smartphone AI development
Talk to us about smartphone AI development
Your development requirements do not need to be fully defined yet. We can help from feasibility checks and PoC discussions to implementation planning.
Your development requirements do not need to be fully defined yet. We can help from feasibility checks and PoC discussions to implementation planning.